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.
Other Government Sources of Information on Immigration It is not surprising that the INS, given its name and notoriety, is often thought of as the one and only source of information on immigration. Certainly, as we have shown in Chapter 4, it is, or could be, the most important--but it is not the only one. This chapter describes the more important of what we have labeled other official sources, those federal agencies and offices which in the performance of their mission come in contact with immigrants or collect information about immigrants. The order of presentation reflects, first, those agencies or offices that have a direct program involvement with immigrants or refugees, followed by those that interact indirectly, those that regularly collect data relevant to immigrants as part of their statistical activities and, finally, those that have the potential for providing useful and important information as a by-product of their activities. As an indication of the pervasive nature of the subject of immigration, this chapter deals with nine separate organizations within six different cabinet-level agencies. We describe briefly both the role of the agency or office and the types of information it collects or produces. (Relevant forms are shown in Appendix A.) In the course of this review, we also comment on weaknesses and makes recommendations for improvements. BUREAU OF CONSULAR AFFAIRS, U. S . DEPARTMENT OF STATE Responsibilities and Data The visa-issuing offices of the State Department abroad are generally the first points of official contact for most aliens wanting either to visit or to migrate to the United States. In fact, the State Department issues two-thirds of all immigrant visas abroad; the remaining one-third go to nonimmigrants and refugees in the United States who adjust to permanent resident status through the INS. Given the current laws and regulations, the Bureau of Consular Affairs exercises enormous discretion in determining who is eligible to receive a visa--as either visitor or immigrant . More important, the State Department is responsible for controlling the numerical limitation on the number of immigrant visas issued. Its actions, in the aggregate, thus can provide the basis for much valuable 74
75 information for monitoring patterns of visa demand worldwide and of movement into the country. Furthermore, since much of the information that subsequently enters the INS data systems originates at this stage, the degree of attention given to full and accurate reporting is crucial. As in most areas in which administrative functions predominate, however, the interest in data and care with which they are collected are very limited. Not only would more care produce better data, but more effective use of the data might also actually improve the administrative functioning. For example, linking visa applications to entry and departure records maintained by INS, which identify possible visa abusers, could provide a basis for improved screening of visa applicants. The statistical programs of the Department of State concerning immigration produce aggregate data on visa applications issued and denied at authorized posts abroad. Reports are prepared at scheduled dates each year showing immigrant visa applications and persons waiting for visas by preference category, foreign state of chargeability, class of visa, and determination (granted or denied) and showing nonimmigrant visa applications by class of visa and outcome. Visas denied are classified by general type of visa and grounds for refusal. The Department of State and INS are collaborating to integrate the process of visa issuance with the process of immigration to the United States by developing a single automated system for recording immigrant visa issuance and utilization. The goal is to monitor efficiently the availability of visas within existing ceilings and to determine the waiting time for applicants by country of chargeability and preference category. The unique information on immigration collected by visa offices, however, concerns events surrounding the process of visa appl ication; that is, the demand for immigrant visas and the way decisions are made by consular officials. These officials administer the various required application forms and also are responsible for considering additional evidence concerning visa eligibility. While files on aliens to whom visas are granted and used are transferred to the INS, which then creates A-files (see Chapter 4), the State Department remains the sole source of information on immigrant and nonimmigrant visa refusals, including the reasons for rejection. The current data reporting procedures do not include characteristics of aliens refused visas, not even country of origin; nor do the data processing systems being developed tap the documentary sources of information available in consular files concerning the process of denying visas to aliens. Traditionally, the Bureau of Consular Affairs has produced data--or program information--primarily for its own use. The data are generally available, although external requests are rare. The bureau has produced annual summaries of data on visa issuance since the early 1920s (although there has been no congressional mandate to do so). Yet it is only now beginning to review what it produces, both in content and format, from the point of view of the administrative uses of the information. The issue of data for analytic or policy use is seldom raised. The State Department clearly has information of immediate and potential value for many analytical purposes, blat it lacks interest in informing potential users of what may be obtained or in making the information available in useful form. The first need, in this as in a number of other areas, is
76 to bring a sense of need and awareness to agency management and to set up much closer coordination in the area of immigration statistics. Recommendations The Panel recommends that the State Department: o Become an active participant in interagency discussions on improving immigration statistics, in order to enhance its own understanding of the need for and uses of such information, and o Establish its own review group to assess the statistical potential of its programs, to propose actions to be taken to improve its statistical performance, and to make recommendations to the department. U.S. DEPARTMENT OF LABOR One of the fundamental concerns surrounding immigration is its impact on the labor force. The U.S. Department of Labor's (DOL) involvement with immigration extends across three of its agencies: the Employment and Training Administration, the Bureau of International Labor Affairs, and the Employment Standards Administration. Within the Employment and Training Administration, two offices deal with immigration to varying degrees: the Division of Foreign Labor Certifications in the U.S. Employment Service and the Office of Research and Evaluation in the Office of Strategic Planning and Policy Development. Division of Foreign Labor Certifications Program and Data The Division of Foreign Labor Certifications is perhaps the best known of the four DOL offices dealing with immigration, since it is involved directly with the actual admission process of certain groups of both temporary and permanent immigrant workers. However, only a relatively small number of immigrants require labor certifications to be admitted to the United States, since current immigration policy is based largely on family reunification. Immigrants entering as immediate relatives outside the preference system or under the relative preferences do not need labor certificates. Only immigrants entering under the occupational preferences (see Table 2-2) or under the nonpreference category (which has been unavailable since late 1978) need the approval of the Division of Foreign Labor Certifications. A total of 54,000 immigrant visa numbers are allotted for the occupational preferences: 27,000 for the third preference category, which includes members of the professions or persons of exceptional ability in the sciences and arts and their families; and 27,000 for the sixth preference category, which includes skilled and unskilled workers in short supply and their families. However, since the numbers allotted include beneficiaries (spouse and unmarried children of the princ ipal immigrant), the actual number of immigrants entering each year with labor
77 certifications is less than half the total number of third and sixth preference visas issued. For example, less than 21,000 ilranigrants were admitted with labor certifications in fiscal 1980, a year in which third and sixth preference admissions exceeded 44,000 and overall immigration exceeded 530,000. The current labor certification program for immigrants was created by the 1965 amendments to the Immigration and Nationality Act. In addition to a provision requiring that aliens seeking admission to the United States as workers have a U.S. job offer, Section 212(a)~14) of the INA requires that each alien intending to migrate to the United States for the purpose of employment must first obtain a certification from the Secretary of Labor that: (1) there are not sufficient U.S. workers who are able, willing, qualified, and available at the place of intended employment and (2) the employment of the alien will not adversely affect the wages and working conditions of U.S. workers similarly employed. When the requirements for the program were first established in the late 1960s, the list of qualifying occupations in each preference category was determined by the division through a more or less systematic inquiry of employment services offices throughout the country. Changes to the list since that time have been both infrequent and ad hoc, reflecting in the main both positive and negative comments from employers on the supply of available U.S. resident applicants for specific occupations. There are two kinds of procedures for securing a labor certification. The more coupon is for the prospective U.S. employer of the alien to apply for certification directly to DOL. The other, far less common, concerns aliens who have obtained U.S. job offers in an occupation precertified by DOL as meeting the requirements of the INA. The Division of Foreign Labor Certifications is relatively small, having a national office staff of approximately 15, 11 of whom are manpower development specialists. However, more personnel are involved in the labor certification program, since applications are processed in approximately 1,900 local offices of state employment service agencies. These applications are forwarded to 10 regional offices (or to the national office, in some instances) for determination. In addition, authority is delegated to the INS and consular officers of the Department of State to grant employer petitions for precertified (Schedule A) occupations. More than 50,000 requests for permanent labor certifications are processed annually. Of these, an average of 25,000 to 30,000 certifications are granted. The Division of Foreign Labor Certifications also has the responsibility for certifying that the admission of temporary agricultural and nonagricultural (H-2 visa) workers will not adversely affect the wages and working conditions of U.S. workers similarly employed and that qualified persons in the U.S. are not available. The determination made by the Department of Labor is advisory; INS makes the final decision regarding the employer's petition for foreign workers. In recent years 15,000 to 20,000 agricultural and logging jobs and an additional 6,000 to 23,000 nonagricultural jobs have been certified annually. However, the number of jobs certified does not indicate the actual number of foreign workers admitted for employment: an employer may use only part, or none, of the certifications granted, and some admitted foreign workers may work in two or more certified jobs.
78 Some statistics from the Office of Foreign Labor Certifications are published annually in the Employment and Training Report of the President. However, the majority of statistical tabulations on labor certifications are unpublished. From 1976 to 1981 approximately nine dif fe rent tabulations were prepared on labor certifications for immigrants. These tabulations included certifications by sex, age, country of birth, intended state of residence, etc. However, beginning in 1982, in order to reduce both the reporting burden and the paperwork, only variables relevant to establishing labor certification have been collected and processed: immigrant certification data are available by determination (approval/denial) cross-classified by detailed occupational codes. These tabulations are available at the national, state, and regional office level. Precertified cases that are handled separately by INS and the State Department are included in the national totals. The variables included in the tabulations of temporary agricultural jobs are number of applications (one application from an employer can cover many jobs), crops or industry (e.g., logging, sheep-raising, etc.), and state of employment. Tabulations by nationality are not possible from DOL records since an employer may import a worker from any nation once the job is certified. Data on temporary nonagricultural jobs are available by determination (approval/denial) and occupational groups. Although limited information is collected by the department on the characteristics of foreign-born workers seeking occupational preferences or admission as H-2 workers, no detail is obtained for employers petitioning for the admissions. Such information would be an important first step toward understanding the demand factors encouraging immigration. For example, a comparison of information on petitioning employers with census data would permit an analysis of the effects of these preferences on the wage rates of native workers. Recommendation The Panel recommends that the U.S. Department of Labor: O Collect and publish summary information on employers who petition for the admission of foreign workers, including those requested under the H-2 provisions. Other DOL Programs The Office of Research and Evaluation in the Office of Strategic Planning and Policy Development is another part of the Employment and Training Administration that has been involved with immigration. Over the past IS years, this office has funded a considerable number of immigration studies, principally on the labor-market characteristics and effects of various kinds of immigration flows. In recent years, however, such funding of immigration research has slowed considerably. The third group involved in immigration.research is the Immigration Policy Group within the Bureau of International Labor Affairs. The group has a staff of four, three of whom are analysts, and conducts analyses of the economic, social , and political aspects of legal immigration (both
79 permanent and temporary), illegal immigration, and refugee movements, both within the United States and internationally. The bureau has also funded some immigration research, on both U.S. and international labor flows. A fourth Labor Department office is involved with immigration as a result of its enforcement efforts. The Wage and Hour Division of the Employment Standards Administration is responsible for enforcing the Fair Labor Standards Act and the Migrant and Seasonal Agricultural Worker Protection Act. Since 1978 this office has carried out a special targeted enforcement program of the Fair Labor Standards Act aimed at employers of undocumented workers. Compliance with minimum wage and hour laws reduces the economic incentive for employers to hire illegal aliens. In fiscal 1983, wage and hour compliance officers investigated more than 18,000 establishments reportedly employing illegal aliens and found more than $40 million in back wages owed to almost 175,000 workers. Statistics other than totals are not tabulated. The Migrant and Seasonal Agricultural Worker Protection Act prohibits farm labor contractors from recruiting, employing, or utilizing with knowledge the services of any alien who is not lawfully admitted for permanent residence or is not authorized by the Attorney General to accept employment. Statistics are available on the total number of farm labor contractors found in violation of the act and the dollar amount of civil money penalty assessments. BUREAU OF REFUGEE PROGRAMS, U.S. DEPARTMENT OF STATE The Bureau of Refugee Programs in the Department of State has three major roles in collecting data related to refugee resettlement: (1) collecting information and reporting on the world refugee situation for purposes of determining U.S. refugee admission numbers and their allocation among regions and refugee groups; (2) collecting and reporting data on refugees admitted to the United States; and (3) monitoring the activities of voluntary agencies that provide reception and placement services to · · arriving refugees. With respect to the first role, the bureau prepares an annual report on the world refugee situation. This report contains both descriptive and statistical information; the latter was presented in a separate volume for the first time in 1984. The statistics focus on two types of refugees: those who are in need of protection and assistance and those who have been resettled and are no longer in need of assistance. The statistics are drawn primarily from information provided by the U.S. embassies and consulates in the countries covered by the report; international organizations such as the U.N. High Commissioner for Refugees and the U.N. Relief and Works Agency, which assists Palestinian refugees; and foreign governments. In addition to the annual report, the bureau publishes data on the refugee situation in Southeast Asia, the area from which the largest number of refugees are admitted to the United States. Each month, the U.S. embassies in this region report the number of land and sea arrivals in countries of first asylum, the number resettled by the United States and other countries, the number transferred to refugee processing centers for training prior to resettlement, and the remaining camp populations.
80 This information is disseminated monthly, by means of summary charts that provide information about each country of first asylum and each refugee group, by nationality, in Southeast Asia. The second role, the collection of information about refugees admitted to the United States, begins with the prescreening of refugee applications, which are used by voluntary agencies under contract to the State Department to determine eligibility for admission. Biographical data are collected about the applicant and family members, including name, sex, birthrate, country of birth, country of residence, marital status, education, occupation, language, and English language capability. These data are transferred to the INS, to be used in making final decisions on admissibility. They are also sent to the Refugee Data Center (RDC) in New York, which is responsible for matching refugee applicants with U.S. sponsors. Under contract with the State Department, the RDC enters these data in its computer system and, using the data bank, determines whether the refugee has ties to any voluntary resettlement agency or has relatives or other close contacts in the United States. This information is used for determining place of resettlement within the United States (for further information on this process, see Chapter 7~. Also under contract with the Bureau of Refugee Programs, the RDC provides monthly reports on expected arrivals of refugees by state and locality. These reports provide biographical information on refugees who have been approved for admission, the agency that will be sponsoring them, and the expected date of their arrival. RDC also collects data on actual arrivals, using statistics from the Intergovernmental Committee for Migration, which arranges transportation for the refugees. Using these data, the Bureau of Refugee Programs publishes monthly reports on arrivals by country of origin. The Bureau of Refugee Programs uses the data it collects for policy-making and administrative purposes. With regard to policy, the data are used to determine the worldwide need for resettlement, in order to determine the U.S. response to this need. Administratively, the bureau attempts to project numbers of future arrivals and to decide whether admissions will be consistent with the number of refugee admissions agreed on as a result of consultation between the President and Congress. This process requires a variety of information, including the number of refugees in first asylum camps, the number of applications pending, the rate at which they are reviewed and the decision reached, the number in language training in refugee processing centers, and the number of recent arrivals. Monthly data provided by the INS are used to regulate the number of new cases that are reviewed and presented to the INS. The object is to avoid large backlogs of approved cases, since the number of refugees who are allowed to enter the United States often changes from year to year. In carrying out its third role--monitoring grants to voluntary agencies to provide reception and placement services--the Bureau of Refugee Programs requires that case files be maintained by voluntary agencies on the services they provide to refugees. These files also contain biographical and other information that is needed to determine what services should be provided. The bureau also conducts site visits to monitor these activities and issues reports on them.
81 OFFICE OF REFUGEE RESETTLEMENT, U. S . DEPARTMENT OF HEALTH AND HUMAN SERVICES The focus of federal refugee program activity is the Office of Refugee Resettlement (ORR). Created by the Refugee Act of 1980, it is located organizationally within the Social Security Administration of the Department of Health and Human Services. In response to the relatively detailed and specific requirements for data collection and reporting to Congress contained in its enabling legislation, the ORR conducts and maintains a number of data collection programs. For example, the primary sources of statistics for the domestic refugee program are the ORR refugee data systems, which contain some 900,000 individual records for refugees (and Cuban-Haitian parolees) of an estimated universe of 1.1 million who have arrived since 1975. At a minimum, each record contains an alien number (assigned by the INS), the year of birth, and the date of arrival. Since 1981 information on country of birth, citizenship, sex, and family and marital status has also been included. These items are obtained from the American Council of Voluntary Agencies, from a form completed during processing in the refugee camp by the sponsoring voluntary agency. Monthly summaries of refugee arrivals are published by the ORR. To the extent possible, the ORR records are updated as further information becomes available. For example, prior to the termination of the Alien Address Registration Program in 1981, information reported by refugees was obtained from the INS and incorporated into the ORR files. As mandated in 1980, INS now serves as the major source of follow-up information, providing the ORR with data obtained from the refugee at the time of application for adjustment to permanent resident status, which can occur from one year after arrival. In addition to basic identifying information, INS collects data on changes in residence since entry, current household composition, education, work history, English language ability, participation in training programs, and the use of cash assistance programs. Finally, additional data elements, such as current occupation and place of residence, are added if and when application is made for naturalization. The office also maintains a system of regular reports by state grantor agencies and compiles selected additional information through agreements with other federal agencies. In addition, it has funded surveys of refugees in order to obtain current information on internal migration, labor market involvement, and other related topics concerning refugee adjustment. (A more detailed discussion of refugee admissions to the United States, illustrating the interactions and interrelationships of the range of participating groups in the process, both outside and in the United States, is presented in Chapter 7, along with additional detail on the activities of ORR.) BUREAU OF THE CENSUS, U. S. DEPARTMENT OF COMMERCE Although not involved in the administration of programs or the implementation of public policy dealing with immigration, the Census Bureau, in carrying out its mandate to produce high-quality statistical data, has a need for and itself produces a wealth of information on
82 immigration. Its need for such information arises from its ongoing program of producing both current and projected national and local-area estimates of the size and characteristics of the population, which requires knowledge of the size and composition of migrant flows into and out of the country. At the same time, the bureau's regular and special programs--the decennial censuses, ongoing sample surveys such as the Current Population Survey, special surveys such as the 1976 Survey of Income and Education, as well as compilations from external administrative systems--all serve as sources of data on the flow of immigrants and emigrants or on the characteristics of the stock of foreign-born persons. Before turning to the detail produced by these diverse sources, it is important to note that the Census Bureau defines immigrants differently than does the INS: the INS is concerned with legal status; the Census Bureau with residence in the United States. To the INS, an immigrant is someone who has been admitted for permanent residence in the United States; people in the United States who entered with nonimmigrant visas, such as students and temporary visitors, are not considered immigrants. Refugees are not counted as immigrants until they adjust their status to that of permanent resident, which can occur one year after entry. In contrast, the Census Bureau defines as immigrants all noncitizens resident in the United States, except temporary visitors and foreign embassy employees who live on embassy grounds. The Census Bureau includes foreign students, temporary workers, refugees, and even illegal aliens as usual residents in its censuses and surveys. In addition to the definitional differences, comparisons between INS data and Census Bureau data are difficult because a census or survey refers to a specific date and registers a status at that time rather than recording the flow of migrants into the country. Death or emigration between a person's entry into the country and the date of a census or survey will preclude counting that person as an immigrant. A census count of immigrants actually represents the balance of persons who immigrated during some prior period less those among them who emigrated or died during the period, rather than all persons who entered the country during the period. The Decennial Census Value and Limitations Information from decennial censuses has two features that make it uniquely valuable in the study of immigrants and immigration. First, the large sample size used to collect detailed data permits statistically precise comparisons of the many different country-of-origin groups at different points in time with respect to a large number of social and economic characteristics, including place of residence, period of entry, citizenship, English language ability, earnings, occupation, labor force status, and home ownership. Second, the historical continuity of the census, and hence relative consistency of census information over time, permit valuable historical comparisons of the status of foreign-born people in the United States.
83 There are, however, a number of important shortcomings of census data for the study of immigration. Because the census serves many important functions (of which the collection of data on immigrants or the foreign-born is only one), the amount of immigration data that can be collected is limited for reasons of both cost and sensitivity. Furthermore, the collection of information about the foreign-born in past censuses has been marked by inconsistency in both the questions asked and the wording used, thus significantly reducing the value of the census as a basis for historical comparisons. Table 5-1 lists the variables of special relevance collected since 1900 on the foreign-born and immigration. As can be seen, such important information as date of entry to the United States, English language ability, and even citizenship have not always been collected. Year of immigration, or the alternate form, number of years in the United States, was asked in the censuses of 1890 through 1930 but not collected in the censuses of 1940, 1950, or 1960. Indeed, 1980 marks the first time since 1930 that two consecutive censuses have collected information on date of entry, permitting comparisons of date-of-entry cohorts over time and thus some assessment of how length of stay in the United States affects social and economic characteristics. Even for this variable, however, the seemingly arbitrary limitation of collection of information on period of entry to only four 5-year intervals for these censuses, with grouped or an open interval for earlier dates of entry, means that only two cohorts of entry (1965-1969, 1960-1964) can be matched and compared. Indeed, apart from place of birth, which has been included as a census item since 1850, none of the special questions relevant to immigration and the foreign-born, on date of entry, prior residence, citizenship, English language ability, and parental nativity, has been collected continuously since 1900. The most consistently collected variable--birthplace of parents--was dropped in the 1980 census, although this item is of key importance in the assessment of the longer-term consequences of immigration and of the legacies of different country-of-origin groups. Each census from 1890 through 1950, and from 1970 through 1980, included questions on citizenship, but in 1960 such questions were asked only in New York State. Nonetheless, the decennial censuses remain one of the most important and valuable sources of data on the foreign-born and their descendants, and particularly so for studies at the regional or state level or for individual cities. As with all such data, those from decennial censuses are subject to various types of errors, including incomplete coverage of the population, errors in reporting, and, in the case of data from a sample of the population, sampling errors. A discussion of the errors and their effects on the results are generally found in the reports presenting the data as well as in separate evaluation reports issued by the Census Bureau. 1980 Census Data Turning to the most recent census, each person in a sample of the 1980 census was asked, "In what State or foreign country was this person born?" Those persons who reported that they were born in a foreign
84 TABLE 5-1 Questions Asked of Special Relevance to Immigration in Selected Censuses Question 1900 1940 1950 1960 1970 1980 Country of birth Yes Yes Yes Yes Yes Yes Year of immigration yesa No No No Yesb YesC Residence 5 years ago No Yes No Yes Yes Yes Citizenship Yes Yes Yes No Yes Yes Ability to speak Yes No No No No Yes English Nativity of parents Yes: Yes Yes Yes Yes No Current occupation No Yes Yes Yes Yes Yes Occupation 5 years ago No No No No Yes No Language spoken at No Yes No No Yes No home when child Education No Yes . Yes Yes Yes Yes Earnings No Yes Yes Yes Yes Yes Note: Includes censuses for which public-use sample tapes are available. aBy s ingle year. bFor intervals: pre-1915, 1915-1924, 1925-1934, 1935-1944, 1945-1949, 1950-1954, 1955-1959, 1960-1964, 1965-1970. CFor intervals: pre-1950, 1950-1959, 1960-1964, 1965-1969, 1970-1974, 1975-1980. country were also asked, "Is this person a naturalized citizen?" and "When did this person come to the United States to stay?" A question on residence five years earlier also was included in the 1980 census sample and was asked about all persons S or more years old. Since this question is directed at all persons in the sample (regardless of nativity), it also serves as an indication of the volume of movement of citizens into the United States from overseas during the previous five years. A similar question was included in the 1970 census. Published volumes by state from the 1980 census show the foreign-born populat ion classif fed by country of birth for the 30 or so countries supplying the largest numbers of foreign-born persons. Other characteristics by which the foreign-bor~ population are classified are year of immigration (5-year intervals from 1960 to 1980, the 10-year interval 1950-1959, and the period prior to 1950) and citizenship. More extensive data are published for all foreign-born and for recent immigrants in a table that shows race, Spanish origin, age, and sex as well as more detailed country listings for year of immigration and
85 citizenship. For persons who immigrated between 1970 and 1980, the following information is tabulated for each state by country of birti household and family characteristics, marital status, fertility, educational attainment, language usage, ability to speak English, employment and occupational characteristics, and income. In addition, the Census Bureau plans to prepare a special report on the foreign-born population, and various computer tape files and special tabulations that provide substantially greater detail will be available. For example, a separate micro-file containing only the records of households containing at least one foreign-born member has been used to produce special tabulations on the characteristics of immigrants, by country of birth. The 1980 census also included a question on ancestry that identified the ethnic origin of all persons regardless of when their ancestors entered the United States. ~ _ ~_ ~ . ~ . Current Population Survey The Current Population Survey (CPS) is a sample survey that has been conducted monthly by the Census Bureau since 1942 in order to provide the official government estimates of national employment and unemployment. Its present sample is about 45,000 interviewed households per month. From time to time, questions dealing with other subjects are added to the survey. Two examples of additions dealing with immigration are provided by the November 1979 and April 1983 surveys. A supplement to the November 1979 CPS was designed to provide basic data on ethnicity and language, to update detailed data obtained in a similar study conducted in 1969, and to bridge the gaps between different measures of ethnicity and language in the 1970 and 1980 censuses: in 1980 questions on ancestry and current language replaced the 1970 questions on country of birth of parents and mother tongue. Questions citizenship, year of immigration, and literacy were also included. In addition, since country of birth of parents was dropped from the 1980 census in favor of self-designated ancestry, the 1979 CPS is the only official source of data on persons of foreign parentage for a recent date. The measurement of ethnicity--indeed the concept itself--is still inexact and previous work has shown that substantial response variability occurs (for example, see Bureau of the Census, 1974~. The 1979 survey permits the examination of the relationship between objective measures of ethnicity, such as country of birth and country of birth of parents, and the subjective measure of self-designated ancestry. These data provide a basis other than the census for examining the characteristics of immigrants and the only recent basis for studying the extent of assimilation of second-generation immigrants. Other characteristics are available from the November 1979 CPS survey: household relationship, age, sex, marital status, veteran status (males only), race, educational attainment, ethnic origin, labor force status, occupation, industry, hours worked, family income, current language, ability to speak English, literacy in any language, literacy in English, and mother tongue. A public-use tape, containing individual records for the 45,000 households in the sample--of which about 2,500 (just over 5 percent) include a foreign-born person--has been released.
86 Although the CPS collects much useful information relevant to studies of immigration, the sample size is too small to support elaborate analysis. In April 1983 a number of questions were included in the CPS to study fertility among foreign-born women. All respondents were asked their place of birth and that of their parents, citizenship status, and date of entry into the United States. Foreign-born women of childbearing age were asked the number of children ever born, the number born outside the United States, and the number currently living in the household. A public-use tape is available, as are tabular results. Again, it is important to note that the relatively small sample size limits the utility of the data. These two surveys are illustrative of the types of questions related to immigration that are included in the CPS from time to time. Other Bureau Sources Some other Census Bureau activities, in a peripheral fashion, are relevant to immigration statistics. One is the Survey of Income and Education, undertaken by the Census Bureau for the then Department of Health, Education, and Welfare in 1976. Included in its extensive data set is information on place of birth, of both the respondents and their parents, English language facility and use, year of immigration, and citizenship status. Both micro-data files and extensive published analyses of the results are available. Another source of immigration-related information is the biennial CPS surveys of voting and registration, which are conducted in national election years. The surveys include questions on voting and registration; their value for immigration-related information lies in the ethnic data and in the reasons for noneligibility for registration, which include noncitizenship. Yet another source of data is the current and post-censal population estimates that are produced by the Census Bureau on a regular basis for a wide variety of geographic and political entities. These estimates include an allowance for net migration (as do official population projections), thus reflecting both immigration and emigration. Although current data on immigration are available from the INS, the movement of people out of the United States is not recorded and, thus, must be estimated by indirect means. For the past 25 years, the Census Bureau has used, in effect, the same annual level of emigration, approximately 40,000, in preparing its current population estimates and projections. Recent evidence, however, suggests a more appropriate level of emigration of legal residents to be around 100,000 per year (Warren and Kraly, 1985~. In light of that evidence, the Census Bureau should reevaluate its use of the smaller number of emigrants per year in its national population estimates program. Until now the Census Bureau has also not included any allowance for net illegal immigration in preparing its population estimates and projections; however, it is currently reviewing its approach (see below). Finally, the ongoing research activities of the Census Bureau present possibilities for obtaining additional information about immigration. Those activities include experimentation with new approaches to estimating emigration and illegal immigration, as well as new and
87 imaginative approaches to sampling that, through cost efficiencies, may permit additional and much needed data gathering on the subject of · · - Immigration. Estimates of Illegal Immigration Illegal immigration presents the Census Bureau with a number of problems that are quite different from those faced by the INS. For example, the Census Bureau is charged by law with counting in its censuses every person whose usual residence is in the United States, regardless of the legality of the person's residence. Thus the Census Bureau's primary current need for separate statistics on illegal immigration is to enable it to assess the completeness of coverage of the 1980 census. In addition, such information is an essential element in improving the accuracy of postcensal population estimates. The immigration figures used in the Census Bureau's programs of current population estimates and projections have not included a specific allowance for net illegal immigration, since illegal immigrants are not included in the data compiled by INS, and it has not been possible to get satisfactory estimates elsewhere. The need for estimates of net illegal immigration in connection with the Census Bureau programs of coverage evaluation and the preparation of current population estimates has led it to continue research aimed at developing such estimates. So far, however, the Census Bureau has not been successful in developing reliable estimates of the number of illegal residents who have entered the United States in any recent period. Some information on the volume of illegal immigration has been developed from the sample portion of the 1980 census. These data proved particularly valuable because many illegal aliens, if not a majority, are believed to have been included in the 1980 census (Warren and Passel, 1983~. There are several reasons for this belief. First, major steps were taken to reduce potential underenumeration, and many illegal aliens could not escape the enumeration net. Second, experimental estimates permitted the inference that, nationally, the population represented by the November 1979 CPS included about 1.25 million illegal aliens who entered the United States between 1970 and 1980. This suggests that the 1980 census, using similar procedures with additional safeguards against omission, must also have included many illegal aliens. Third, the 1980 census counted nearly 6 million more people than were expected in 1980 on the basis of the 1970 census and 8SO,OOO more people than were estimated to be residing legally in the United States in 1980. Using the data from the sample portion of the 1980 census, the Census Bureau estimated that about 2 million illegal aliens were included in the census, of whom approximately half were born in Mexico. The estimates were derived by comparing the census count of aliens with estimates of the legally resident alien population, based on data collected by the INS in January 1980. Although the illegal alien population may not be accurately measurable, these results will assist in placing some reasonable bounds on its size. The panel commends the bureau for its innovative work in this area and strongly urges that it be continued. A brief review of methods used to estimate the size of the illegal alien population, including the Census Bureau's method, appears in Appendix B.
88 Although all these studies suffer from uncertainties, it seems likely that there were between 2 and 4 million illegal aliens in the United States in 1980, and that the number has not been growing remarkably fast In recent years. Recommendations The Panel recommends that the Census Bureau: o Restore to the 1990 census a question on place of birth of parents. Collected in each census from 1900 through 1970, this information is key to the study of assimilation by immigrants and their children; 0 Make every effort to maintain consistency in question wording and tabulation detail from one decennial census to the next in areas relating to the foreign-born, so that immigrant cohorts recorded in successive censuses can be linked or compared on a consistent and common basis in order to assess adjustment experience; o Initiate a program of research leading to improved estimates of emigration. As one approach, the panel strongly endorses testing, as a supplement to the CPS, an emigration module involving the collection of information on the residence of close relatives. TRANSPORTATION SYSTEMS CENTER, U.S. DEPARTMENT OF TRANSPORTATION The Transportation Systems Center (TSC) of the Department of Transportation monitors international arrivals and departures by air through two independent studies, one based on the INS form I-92, which summarizes arrivals and departures by flight, and the other, an in-flight survey of passengers conducted by participating airlines. International Arrivals and Departures by Air The TSC established a program for processing information from the INS form I-92 in 1972 under a joint working relationship between the Departments of Transportation, Commerce, and State, the INS, and the Civil Aeronautics Board. The object of the program is to produce statistics on international arrivals and departures by air, excluding those between the United States and Canada. The data are intended for use by transportation policy planners and analysts, primarily in international negotiations over carrier market shares, but they also can be used for balance of payments estimates, estimates of net migration to the United States by air, and studies of tourism. Figures are published in monthly, quarterly, and annual reports, and monthly tape summaries are also available. A complete data base is maintained beginning with 1976. Although the purpose of the program is not the measurement of immigration, but rather all arrivals and departures by air, the program is discussed in some detail here because it represents a clear example of how a lack of interagency cooperation can disrupt a useful statistical system.
89 The basis of this statistical system is the INS form I-92. The completion of this form is required for each arriving or departing international flight except for direct flights to or from Canada. The form is completed by airline staff and collects the following information: last foreign port for arrivals or first foreign port for departures; airline; flight number; port (arrival or departure); date; type of transport (scheduled commercial, charter commercial, U.S. military, foreign military); and passengers by port of departure (arrivals) or destination (departures) by citizens, noncitizens, and total. Airline personnel provide the count of total passengers by port; the count of noncitizens used to be obtained from summing the INS I-94 forms (for nonimmigrants; see Chapter 4) for the flight, and the count of citizens was obtained as the residual (total passengers less noncitizens). Prior to 1983 the I-92 forms were sent daily to the TSC processing contractor after some verification by INS inspectors against the corresponding I-94 foes; the extent of this quality control apparently varied from airport to airport. In January 1983 the INS, without consultation and in contravention of a memorandum of understanding with the TSC, changed both the I-94 form and the process for completing and processing the I-92 form. The question on port of embarkation (arrivals) or disembarkation (departures) was dropped from the revised I-94, thus eliminating a consistency check for noncitizens between the I-92 and the corresponding I-94s. At the same time, returning, first arriving, or departing resident aliens were no longer required to complete I-94 forms, thus removing the basis for obtaining the overall citizen-noncitizen breakdown for each flight. The INS also instructed airlines that they need no longer complete the total passenger column for each airport on multistop flights, which represent about 40 percent of international flights serving the United States. The process was changed so that INS inspectors no longer verify the I-92 form against the corresponding I-94 forms. The I-92 is bundled with the I-94s for the flight and sent to the INS processing center, where some consistency checks are supposed to be carried out between the I-92 and I-94s. However, the number of possible checks is small (for example, the number of aliens on the I-92 cannot be smaller than the number of I-94s, though it can, and often will, be larger). The INS processing contractor is then responsible for sending the I-92s to the TSC contractor. The changes introduced by INS in January 1983 have reduced the potential for consistency checks, increased the risk of loss of forms, introduced substantial delays into the document flow, and, most important, made it impossible to obtain citizen-noncitizen breakdowns by airport on multistop flights. This last issue deserves some further clarification. Through 1982, I-94 forms for arriving or departing aliens were used to obtain the number of aliens by port of origin or port of destination. The I-92 gave the total number of passengers by port of origin or destination, and the number of U.S. citizens by port of origin or destination was then obtained by subtraction; failure to turn in I-94s on departure would therefore inflate the number of departing citizens and reduce the number of departing aliens. The INS changes affect this procedure in three ways: the I-92 no longer records total passengers by port, the I-94s no longer record port for noncitizens, and I-94s are not required of permanent resident aliens or arriving permanent immigrants.
JO The first two changes affect only multistop flights, but the third change affects all flights. The INS changes have thus greatly reduced the detail available in the data (the TSC is estimating the breakdowns on the basis of past patterns) and have jeopardized a useful data set covering the period 1976-1982. Negotiations are still continuing between the TSC and the INS about this s i tuat ion . There are some doubts about how reliable the information produced by the system has been. Vining ( 1980) evaluated data quality in a study of net immigration by air, using the balance of arrivals over departures as an indicator o f a lower bound for tote 1 net immigration. He compared TSC data on arrivals and departures on North Atlantic routes with International Air Transport Association (IATA) data, and he compared data on all arrivals and departures with figures on international air passenger traffic to and from most major international airports in the United States. Though these other data sets are not strictly comparable with the TSC figures--IATA excludes passengers on nonmember airlines and some charters, and the airport data generally include movements to and from Canada--all three show broadly similar trends over time in arrivals, departures, and net arrivals. However, net arrivals as a proportion of total traffic are almost twice as high for the TSC data as for either the IATA or airport data, the differences being accounted for by very high TSC net arrival rates for charter flights. Vining suggests that the discrepancy arises from the failure by some departing charter flights to complete I-92 forms, thus underreporting total departures by air, but argues that the consistency between the three data sources in other respects supports the reliability of the TSC data at least for scheduled flights. A comparison of the TSC and airport data on an airport-by-airport basis (given by Vining in an appendix) is not reassuring, however. Although arrivals and departures from the two sources do follow broadly similar trends, the sizes of the flows are often substantially different. These differences can be partly explained by the fact that the universes are not the same--airport data typically include flows to and from Canada and in some cases other places not included in the TSC data--but not entirely. The most dramatic differences appear for Anchorage: for 1972-1978, TSC annual arrivals average 167,000 and departures 240,000, while the airport data, which include Canadian traffic, average 40,000 for both arrivals and departures. If the elements can include discrepancies of this order of magnitude, one cannot have much confidence in the totals or the net balances. An examination of the internal detail of the TSC tabulations suggests that Vining is correct about an undercoverage by the I-92 of departing charter flights and also that departing aliens are sometimes reported as departing citizens, inflating the number of the latter at the expense of the number of the former. A plausible explanation for this discrepancy is a failure to collect the departure sections of the I-94 forms from all departing aliens. The TSC data for 1982 for scheduled flights show an excess of citizen departures over citizen arrivals of 1.166 million and an excess of alien arrivals over alien departures of 1.731 million. Both net flows are unreasonably large, probably because 5-10 percent of departing aliens are reported as departing citizens. For charter flights, arriving citizens exceed departing citizens by 198,000, and arriving aliens outnumber departing aliens by 91,000. The net inflow of
91 citizens by charter flights, in contrast to the net outflow by scheduled flights, suggests undercoverage of departing flights, an undercoverage that seems to be more serious for U.S. flag carriers than for foreign flag carriers. The TSC data for international arrival and departure by air represent valuable information on travel to and from the United States. The data are flawed, however, particularly with respect to coverage and classification by citizenship of departures. No direct checks on arrivals data are possible, but in view of INS procedures these figures are probably reasonably reliable. The data are useful for market share negotiations and studies of tourism, although attempts to estimate net migration flow from them are probably inappropriate. The changes in the I-94 form and procedures introduced by INS in 1983 disrupted the processing system, introducing delays and greatly reducing the amount of detail available for the system. The situation could be improved in at least two ways. If airlines completed the I-92 forms fully--citizens, aliens, and total passengers boarding or alighting at each port--then I-94 forms would not be needed at all. The airlines' computer booking systems could probably provide the I-92 information without difficulty. Alternatively, the I-94 form could be modified to collect information on port of boarding or destination, with airline personnel recording total passengers by port, and the original system could be continued. It might not be necessary to return to requiring permanent residents to complete I-94s, since for most purposes a division of traffic into two categories--citizens plus permanent residents and other aliens--would be closer to what analysts want (a residence classification) than a citizen-noncitizen split. An effort should also be made by the INS to improve the coverage and quality of data on departures, to increase the overall value and applicability of the data. Survey of International Air Travelers The Travel and Tourism Administration of the U.S. Department of Commerce is mandated to develop the necessary statistical and market research data on international travel to facilitate and guide planning in the public and private sectors. To meet this requirement, an in-flight survey of international departing travelers was established late in 1982. The survey planning and execution are carried out by the TSC under an interagency agreement; the program is also sponsored by the National Park Service. Participation in the survey is voluntary for both airlines and passengers. At the end of the first quarter of 1983, 32 of the 53 major international airlines regarded as of special significance for the program were participating, although some major carriers, among them British Airways, were not. A stratified random sample of scheduled flights of participating airlines is drawn from the Official Airline Guide covering the week following the third Monday of each month; the strata are airlines or, for U.S. carriers, airline-route groupings. All passengers on selected flights are included in the sample, and those who respond are assumed to represent all passengers. The four-page questionnaire collects information on place of origin and destination; lifetime air trips to or from the United States; with whom traveling;
92 purpose of trip; type of ticket; source of information and bookings for trip; age; sex; occupation; household income; citizenship; residence; for visitors to the United States, port of entry, length of stay, places visited, and expenditures by category; and for U.S. residents, length of trip, places to be visited, and expected costs. The questionnaires are distributed by airline personnel at the beginning of the flight and collected before disembarkation. Statistics are generated from the responses by using additional information from the TSC data on international departures by air derived from I-92 forms. Responses on a particular flight are first weighted to allow for children in the respondent's party, then weighted to allow for nonresponse on the flight, then weighted to represent all travelers on all flights in the stratum, and finally weighted using I-92 data by residence and port of destination to approximate traffic of nonparticipating as well as participating airlines. The most serious problem with the survey is the response rate, of both airline and traveller. In the first quarter of 1983, only 60 percent of key airlines participated; for these, the flight response rate--flights for which completed questionnaires were returned for processing--was 51 percent (266 of 524~; and the average passenger response rate for those flights with responses was about 31 percent. Thus only about 16 percent of the intended sample provided responses, and the intended sample excludes an unspecified number of travelers on the airlines that are not participating in the survey. It is doubtful in these circumstances whether the data can be regarded as better than broadly indicative of the characteristics of departing travelers. The Survey of International Air Travel is not strictly relevant to a study of migration statistics, since its primary focus is on tourism. It has been discussed here because in theory it covers all departing passengers and uses INS information for data-weighting purposes. The low response rate raises the question of whether the objectives of the survey could be better served by interviews of a smaller sample of departing passengers (who usually have a few spare minutes between check-in and boarding) than by a large sample with a poor response rate depending on self-completion and airline cooperation. Since we have not had the opportunity to examine all the issues involved, the panel recommends that: o The Travel and Tourism Administration explore alternatives to the present approach with a view toward obtaining more reliable information. SOCIAL SECURITY ADMINISTRATION, U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Programs and Data The Social Security Administration (SSA) is a largely unexploited yet potentially valuable source of information for analyses of immigration to the United States. On one hand, SSA has a highly professional, well-managed statistical operation that, in supporting the administration of social security programs, has developed data sets covering a
93 substantial proportion of the foreign-born population, both aliens and naturalized citizens, resident in the United States. Social security data thus are potentially useful for analyses of the characteristics of immigrant populat ions and for studying the impac t of immigrate ion on domestic sectors. On the other hand, in seeking access to social security data on migrants, an analyst faces formidable obstacles that derive from policies concerning the availability of federal statistics and from the distribution of resources and ordering of priorities within SSA. The general trend over the past few years has been to impose relatively greater restrictions on the accessibility of SSA data to researchers both from other government agencies and from outside government. The potential usefulness of SSA data systems for research on international migration derives both from the characteristics of populations covered by social security programs and from the specific items of information that are recorded. Basically, two populations are covered in SSA administrative files: workers whose employment is covered by social security programs and persons receiving benefits from one or more of the social security programs. Approximately 90 percent of U.S. workers, including the self-employed, are covered by the social security system; the major groups excluded are federal employees and persons whose earnings fall below a specified level in a given quarter. Benefits are paid under several social insurance programs, such as Old-Age, Survivors, and Disability Insurance (OASDI) and Supplemental Security Income (SSI). The administration of social security programs sometimes identifies populations of international migrants. The most obvious example is that of the population of social security beneficiaries residing abroad. Special requirements for awarding benefits outside the country have resulted in an administrative data set for beneficiaries residing in foreign countries, most of whom are foreign-born, both retired workers and elderly dependents. These data have been used by demographers over the years as indicators of the process of emigration from the United States and return migration of former immigrants. Social security program requirements have also generated statistical documentation of the social and economic status of recent refugee groups admitted to the United States. The concern of several of these analyses has been to determine the patterns of use over time of Supplemental Security Income among refugee populations (see Kahn, 1980; Merritt, 1982; Grossman, 1978~. The 1973 exact match study is an illustration of how information on international migration can be generated as a by-product of evaluation of SSA programs and policies. The study was undertaken to document patterns of income distribution among persons and families by matching SSA records on reported earnings and benefits with IRS tax records. However, the final matched files also were used in the application of an estimation technique to produce a measure of the illegal alien population resident in the country (Lancaster and Scheuren, 1977~. The exact match study also illustrates the important role of SSA administrative data sets in efforts to link components of the federal statistical system. Several indicators of migration status are maintained in the micro-data files supporting the administration of social security programs; as a result, these files have substantial inherent potential for analyses of immigrant populations. For example, all applications for a social security number (form SS-5) filed since the initiation of the
94 program have been coded by SSA to create the NUMIDENT file. Country of birth is coded for foreign-born persons ~ s Late of birth for the native-born); date of application, age, sex, and race are also coded. Date of application can be used as a proxy for period of arrival in the United States. The file can be linked with other SSA files on earnings and benefits for individual workers and beneficiaries. Analyses of earnings, industry, geographic mobility, disability, and types of benefits can thus be conducted for a range of social and demographic groups, including immigrant cohorts. The social security application form was revised in 1980 to record the citizenship and visa status of the applicant; thus, for recent applicants the NUMIDENT file now includes four classifications of legal status: U. S. citizen, legal alien allowed to work, legal alien not allowed to work, and other. These categories make possible comparisons of certain economic characteristics between native U. S . cit izens, naturalized U.S. citizens (combining information on place of birth with citizenship status), and the alien groups. (The instructions for completing forts SS-5 do not identify the specific visa categories for which employment in the United States is authorized. The category "legal alien allowed to work" will include both permanent resident aliens and certain classes of nonimmigrants, such as temporary workers, intracompany transferees, and some foreign students.) District offices of SSA do in fact produce monthly tabulations of applicants for social security numbers by citizenship status, race, and age. The potential of the SSA system to generate policy-relevant analyses of populations of international migrants has already been demonstrated through the exact match study and the various projects concerning the characteristics of specific refugee groups covered in SSA programs. Unfortunately, obstacles to the use of SSA data on immigrant populations exist throughout the SSA statistical system. The quality of data on immigrant status is one such obstacle, since country of birth is missing from approximately one-fifth of the records. The keying of data from form SS-5 is now more complete for new applicants, but the validity of the information recorded on the application form and accuracy of keying in the SSA distric t of fices has not been as sessed. The SSA has made little analytical use of its data on international migration; the agency's program and priorities have not generally required evaluation of the benefit and earnings characteristics of foreign-born groups, although recent refugee admissions are a notable exception. Releasing data to outside users should result in more extensive use, but two impediments to such release exist. First, SSA currently lacks sufficient resources throughout its data processing and analytic units to support an ef feet ive program of user service, even on a cost-reimbursable basis. Furthermore, there is a basis for skepticism concerning the agency's current commitment to user service: procedures for making data available are not standardized; policy is not clear as to what services can be supplied, that is, purchased by outside users; and documentation of data sets is e lusive . The second impediment concerns the issue of confidentiality. Until relatively recently, SSA, through its Of f ic e of Research and S tat i s t ic s , maintained a program of data user service that set an example for the rest of the federal statistical system. The star in the program was the Continuous Work History Sample (CWHS), supported in a manner to ensure
95 data quality, to maintain and update files, deco provide documentation and user outreach, and even to provide incentives for outside use through a program of research grants. However, the CWHS has not been available for public use since 197S, at least partly because of constraints on the dissemination of these data resulting from interpretations of the Privacy Act of 1974; SSA no longer disseminates administrative files for outside research. The restrictions imposed on SSA data have affected many areas of social science research, including research on immigration. It is unfortunate that this change in policy occurred when the potential for analysis actually increased, with the revision of SSA documentary resources and when other data sources on the immigrant population (in particular, the INS's alien address report program) were eliminated. The statistical information maintained by SSA is of considerable potential value for analyses of U.S. immigration. Recommendations The panel recommends that the Social Security Administration: 0 Incorporate data on country of birth and, when possible, citizenship and visa status into its series of tabulations on workers and beneficiaries; 0 Develop a program of special tabulations to be available on a cost-reimbursable basis to support analyses of the economic and social integration of immigrant groups; 0 Develop appropriate mechanisms for evaluating and, if necessary, Improving the quality of data on country of birth and legal residence status that are captured in its statistical systems; o Explore with the Statistical Policy Office of the U.S. Office of Management and Budget approaches to permit the sharing of CWHS files with researchers and analysts outside the agency. NATIONAL CENTER FOR HEALTH STATISTICS' U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES The National Center for Health Statistics (NCHS) is responsible for the collection, analysis, and dissemination of statistics on vital events in the United States: that is, births, deaths, marriages, and divorces. The collection of vital statistics in the United States is decentralized. Events are first registered at the local level by state of occurrence; NCHS subsequently also produces tabulations by place of residence of the mother. Certificates for out-of-state-of-residence events are forwarded to the state of residence for processing and tabulation. Under a federal-state cooperative arrangement, the states then forward information to NCHS for processing of national birth and death data. It should be noted that, although all states participate in collecting data on births and deaths, the cooperative arrangement reserves a great deal of flexibility and independent action to the states. Currently, between 44 and 46 states submit vital statistics data
96 to NCHS in the form of computer tapes; the remaining states send microfilm copies of vital statistics certificates. The standard vital statistics records of birth, death, marriage, and divorce certificates contain information on the place of birth of the personas) and on place of residence. Thus, there is the potential to tabulate vital events for the foreign-born populat ion. Although the item on country of birth is coded, the detail, unfortunately, is limited to state in the United States, Puerto Rico, Virgin Islands, Guam, Canada, Cuba, Mexico, remainder of the world, and place not stated. An item on the ethnic origin of the parents of a live birth was included in the 1978 revision of the standard birth certificate and has been incorporated on the birth record of 22 states. The item on place of residence specifies a code for nonresidents of the United States, but the criteria for determining residence are not specified in available documentation. Vital statistics are published in a variety of NCHS publications. Data on nativity, however, have been included in annual volumes only intermittently. Annual volumes do include a tabulation for births and deaths cross-classified by state of occurrence and state of residence; the residual refers to nonresidents of the U.S. and is an interesting indicator of the utilization of health facilities in southwestern states by aliens. Special reports on the fertility of specific parentage groups have appeared in recent years; however, published analyses and evaluation of trends in marriage and mortality are sorely deficient. A program of public-use research files is promoted by NCHS. Annual micro-data files are available for births, deaths, marriages, and divorces. However, place of birth is coded as described above in the first three of these files, so use of the information on nativity is quite limited. The vital registration system of the United States does have the potential of producing annual tallies of most vital events occurring to aliens living in the country, and infest ion on vital events occurring to immigrants has substantial potential for studies and policy on immigrat ion. Data on births occurring to immigrants would provide a direct basis for calculating their fertility rates, useful for studying the impact of immigration on population growth and the pattern of adjustment of it igrant fertility to that of the established population. Similarly, information on deaths of immigrants would permit the direct study of immigrant mortality, would permit the INS to keep its record-keeping systems more current, and could potentially provide a basis for estimating a lower bound for the size of the population of illegal immigrants. It is evident from this review, however, that the immigration data collected through the vital registration system remain remarkably underexploited. One reason for the low degree of analysis of nativity data is the lack of confidence in its quality, particularly for the foreign-born population. Given the importance of documenting the demographic experience of immigrant groups and the general role of immigration in national populat ion dynamics, the NCHS should improve its data.
97 Recommendations The Panel recommends that the National Center for Health Statistics: o Initiate the coding of the specific country of birth for vital events; o Routinely prepare analyses and publish tabulations of vital events for immigrant populations; o Institute a program to evaluate and improve the quality of data on place of birth. CUSTOMS SERVICE, U.S. DEPARTMENT OF THE TREASURY Data Collected Travelers arriving in the United States by air or sea, except those arriving on direct flights from Canada, are required to complete the Customs Service declaration card, form 6059B; one form is completed per arriving family group. Arrivals across land borders are not in general required to complete the form, unless referred to secondary inspection. The information collected is name, date of birth, airline and flight number, U.S. address, citizenship, whether permanently resident in the United States, and if not, expected length of stay, purpose of visit (business or pleasure), and items pertaining specifically to Customs Service activities. The form used until mid-1983 also collected the number of people in the family group, but this question was excluded from the revised form. In addition to a form for each family group, the Customs Service also uses a workload report, form 5905, that summarizes the total number of passengers and crew, the total number of declarations by passengers and crew, and the total number of baggage examinations for each flight. The data collected by form 5905 are thus very similar to those collected by the INS form I-92, though without any citizenship or residence status breakdown. No statistical use has ever been made of the information collected by the customs forms. The family group forms are packaged by flight number and stored in boxes for a period of two years, after which they are thrown out. The only use ever made of the forms after inspection is in particular investigations, to trace the movements and declarations of individuals over the preceding two years by making a hand search of the forms boxed by flight number and date. As a potential source of information about immigration, the Customs Service form 6059B has some important similarities and some important dissimilarities to the INS I-94; some are positive features, some negative. Unlike the I-94, the customs form is filled in regardless of citizenship or residence and thus could provide information about numbers of returning citizens. On the negative side, however, only one form is filled in per family group, with no record on the new form of the number of people in the group, so it does not provide directly the total number of people arriving. (The summary form 5905 gives the average number of members in each family group for each flight as the ratio of total passengers to total declarations, but the average might vary systematically by citizenship and thus not give a very accurate expansion
98 factor for obtaining total arriving U. S. citizens from total arriving family groups headed by a U.S. citizen.) On the positive side, the form records permanent residence (in the United States or elsewhere) and could thus distinguish between nonresident and resident U.S. citizens--although that information is never used, and we do not see how it could be used to learn anything about U.S. citizens living outside the country. Neither positive nor negative, relative to the I-94, is the limitation of coverage of the 6059B form to airports and seaports: the vast majority of entries to the United States occur at land ports of entry, across the Canadian and Mexican borders, and only a small fraction of such entries will result in a completed customs form. Thus from the point of view of measuring the gross flow of people into the United States, the customs form does not offer information of sufficient value to justify any data processing for statistical use. Recommendation The Panel recommends that : o The Customs Service and the INS jointly consider merging the two forms currently completed by arriving passengers. The similarities between the Customs Service form 6059B and the INS I-94 are such that the existence of two separate forms is not justified. The only information not common to both forms is, on the I-94, the city in which the alien's visa was issued, country of residence, and the departure record, and on the 6059B, purpose of trip, expected length of stay, and declarations concerning agricultural regulations, currency, and goods obtained abroad. The 60598 is completed by all heads of family groups, the I-94 by all aliens except permanent residents. A single form ful fit ling both purposes could be required for all arriving passengers without any great increase in the paperwork burden on passengers, and the machine processing of the information by the new nonimmigrant information system would surely offer advantages to the Customs Service for checking on the movements of individuals under surveillance, as well as offering the potential for recording accurately, with characteristics of statistical interest such as age and country of residence, the numbers of people entering the United States through air and sea ports of entry. The potential gain from such a merger would seem to just if y the relatively modest costs involved. GOVERNMENTWIDE ISSUES The foregoing sections clearly illustrate the extent of involvement of the federal government, outside the INS, in collecting or producing information relevant to the study and understanding of immigration. Some of the agencies, such as the Bureau of Consular Affairs in the State Department, deal with the process of becoming an immigrant and have data sets offering insight into the selection, numbers, and movements of persons . Others , such as the Social Security Administration, deal with concerns or developments after arrival, which give insight into the processes of adjustment. Together with the INS, they produce or have the
99 potential to produce a broad spectrum of information relevant to policy makers. Because the nine agencies covered have such different functions and objectives, we chose to present recommendations focused on each agency's operations or activities within the context of our discussion of that agency. In its review, however, the panel noted a number of issues that concern all the agencies, including the INS. This section is therefore addressed to all the government agencies that have a role in providing information on immigration and emigration and to the agency with oversight responsibility for all of them. The Office of Management and Budget, through its review of budget proposals as well as by statute, plays a key role in establishing the nation's statistical agenda and in monitoring its progress. The panel believes that in recent years these functions have not been exercised appropriately and, consequently, data quality has suffered, coordination among agencies has deteriorated, timeliness has been ignored, and, in general, efforts to improve statistics on emigration and immigration have foundered for lack of leadership and concern. Certainly, each agency must be responsible for its own activities, but OMB, given its position in the executive office of the President, has an overriding ability and responsibility to act as a catalyst and to initiate the process leading to significant improvements in the data base. The panel thus strongly recommends that the director of the Office of Management and Budget: o Ensure that OMB exercises its responsibilities to monitor and review statistical activities and budgets concerning immigration statistics, particularly those of the INS, to minimize duplication and make sure that appropriate procedures are used, standards met, and priorities observed in the collection, production, and publication of such data; o Require and establish the mechanisms for continuing interagency coordination in the field of immigration data; participate in discussions designed to achieve consistency and comparability in concepts and definitions used by the relevant agencies in the collection of such information; and oversee the introduction and use of standardized approaches; 0 Actively encourage and monitor the timely publication and dissemination of immigration data; and 0 Actively encourage and support the preparation and release of micro-level public-use tapes conforming to the requirements of the Privacy Act and accompanied by adequate documentation, for use by researchers. A common thread during discussions with virtually everyone concerned with immigration statistics was the obvious lack of coordination between the many agencies involved and their seeming indifference to users, as illustrated by the lack of timeliness in releasing data, failure to address users' needs or apply appropriate quality control standards, and inconsistencies in definitions.
100 The panel therefore further recommends that the respective agencies undertake a review of the ir data-gathering ef forts in the area of ir~igrat ion, in order to: c, Minimize duplication and burden and maximize quality and utility of the collected information; o Develop approaches leading to the timely publication and dissemination of such data, including, as appropriate, the preparation and release of fully documented micro-data, public-use tapes; and 0 Establish and maintain formal liaison with other federal agencies involved in the collection or analysis of immigration related data. REFERENCES Bureau of the Census .1974 Consistency of Reporting of Ethnic Origin in the Current Population Survey, Technical Paper Number 31, February 1974. Grossman, H.A. 1978 OASDHI-Covered Earnings of Indochina Refugees, 1975, Social Security Bulletin (41, 6) Social Security Bulletin. Kahn, A.L. 1980 Indochina Refugee Receiving Supplemental Security Income, July 197B, Research and Statistics Note (No. 6) Social Security l Administrat ion. Keely, C. ~ and Kraly, E. 1978 Recent Net Alien Immigration to the U. S.: Its Impac t on Popul at ion Growth and Native Fertility. Demography 15~3) :267-284. Lancaster, C., and Scheuren, F.J. 1977 Counting the Uncountable Illegals: Some Initial Statistical Speculations Employing Capture-Recapture Techniques. Paper presented at the Annual Meeting of the American Statistical Association. Merritt, M.G. 1982 Cuban and Haitian Applicants for SSI Disability Payments. Social Security Administration mimeo. Vining, D.R. 1980 Net Migration by Air: A Lower Bound on Total Net Migration to the United States, Working Papers in Regional Science and Transportation, University of Pennsylvania. Warren, R., and Passel, J. 1983 Estimates of Illegal Aliens from Mexico Counted in the 1980 United States Census. Paper presented at the Annual Meeting of the Population Association of America, Pittsburgh, Pennsylvania.