From Foresight, No. 47
published 2006
Reliable data is a necessary ingredient for credible policy and its implementation.
U.S. Commission on Immigration Reform—1994.
Immigration policy once again stands at the forefront of the national political agenda. The National Conference of State Legislatures reports that almost 500 pieces of legislation—dealing with everything from using public services like health care and education to strengthening penalties for offenses like human trafficking and hiring illegal aliens—were introduced in state legislatures this year.(1) However, because immigration data are so sketchy, the U.S. General Accounting Office has questioned its usefulness(2) and the Congressional Budget Office says that the unauthorized immigrant population can be estimated “only imprecisely.”(3) Nonetheless, policymakers must craft immigration policy despite the considerable uncertainty surrounding the size and location of the country’s illegal immigrant population.(4)
Here we present an alternative approach for estimating the size of Kentucky’s illegal immigrant population. Accurate, scientifically based estimates of the illegal immigrant population, rather than anecdotal speculation, are the only reliable foundation for sound and appropriate public policy in this area. Rather than scattershot responses to conjecture, they enable the development of focused, realistic, and cost-effective public policy. This approach not only allows one to generate size estimates of the undocumented population, but also creates a framework within which one can evaluate the plausibility of estimates from governmental agencies and research organizations. Our analysis indicates that Kentucky’s undocumented population has been growing steadily and currently stands at about 26,000, which is lower than most other estimates bandied about.
Accurate population numbers are important since virtually all state and local government planning and spending are predicated, to some degree, on population size and location. Education, health care, and transportation expenditures, for instance, are directly affected by the number of students, patients, and commuters.(5) Moreover, any new federal legislation affecting the status of illegal immigrants would likely have financial implications for state and local governments. The financial effect, of course, depends on the exact contours of any proposed legislation, but illegal immigrants could become eligible for a range of public assistance programs.(6) Consequently, good data on the immigrant population is as important today as it was over a decade ago when the U.S. Commission on Immigration Reform wrote “reliable data is a necessary ingredient for credible policy and its implementation.”(7)
The estimated number of illegal immigrants in the United States ranges from about 12 million to as high as 20 million.(8) The 12 million person estimate is probably the most frequently cited and is derived from the so-called “residual method.”(9) Essentially, this method assumes that the Census Bureau accounts for most of the foreign-born residents in this country, both legal and illegal.(10) The estimated number of legal immigrants, which is derived from multiple official government sources, is subtracted from the total number of foreign-born. The resulting “residual” is the estimated undocumented or illegal immigrant population.
Researchers using this method have estimated that Kentucky’s illegal or undocumented population was between 30,000 and 60,000 in 2005.(11) Figure 1 shows the various published estimates of Kentucky’s illegal population from 1990 to 2005. The U.S. Immigration and Naturalization Service (INS) estimate for Kentucky in 2000, as well as the Pew estimates for 2005 and 2006, were generated using the residual method.(12)
Figure 1: Estimates of Kentucky’s Undocumented Population
However, some have disputed the notion that the Census Bureau accounts for most of illegal immigrants in its foreign-born estimate. For example, using a less-than-transparent method that looks at the growth of financial remittances to Mexico, housing permits in so-called gateway communities, school enrollments, and cross-border flows, the investment banking firm of Bear Stearns issued a report in 2005 asserting that “the number of illegal immigrants in the United States may be as high as 20 million people, more than double the official 9 million people estimated by the Census Bureau.”(13)
It is certainly possible that the estimated 30,000 to 60,000 undocumented individuals in Kentucky, while derived using a rigorous method and official government data, is too low. Indeed, this seems to be the conventional wisdom. For example, the Kentucky State Data Center estimates that Kentucky’s Hispanic population is three to four times larger than the official Census estimate of about 75,000.(14) Moreover, the Lexington-Fayette County Urban Government released a report in December 2005 contending that “anecdotal information suggests that more than 90 percent of the Hispanic population in Central Kentucky is undocumented,”(15) while an agricultural sociologist at Western Kentucky University has asserted that “only about 15 percent of the Latinos now living in Kentucky have legal immigration status.”(16)
These higher estimates could be correct, but there is no way to determine their accuracy based on the soundness of the assumptions, quality of the data, or rigor of the statistical technique because they are based on anecdotal information or are generated using less-than-transparent methods. Therefore, we have developed a transparent approach that uses data from multiple sources to assess the internal consistency and plausibility of related factors and estimates of Kentucky’s undocumented population.
We establish likely values for three factors: 1) the estimated percentage of Limited English Proficient (LEP) students who are undocumented or illegal;(17) 2) the estimated percentage of the undocumented school-age children enrolled in school; and 3) the estimated percentage of the total undocumented population who are undocumented school-age children. Then, we generate estimates of the total undocumented population using these factors and can assess their veracity based on the values used for the three basic factors.
The following example illustrates the approach. We know there are approximately 11,200 LEP students in Kentucky’s public schools. We assume that approximately 15 percent are undocumented, resulting in 1,680 undocumented students (11,200×0.15=1,680). It is unlikely this figure represents all of the undocumented school-age children in Kentucky since not all are likely to attend school. If we assume that 82 percent are enrolled in school, then there would be a total of 2,049 undocumented school-age children living in Kentucky (1,680÷0.82=2,049). Finally, if we assume that the 2,049 undocumented school-age children constitute 8 percent of the total undocumented population, then Kentucky’s undocumented population would be nearly 26,000 (2,049÷0.08=25,613).(18) We can then assess the likelihood of the 26,000 person estimate based on plausibility of the three factors used to generate it. In the sections below we explain how we derived the expected values for these three factors.
Limited English Proficient Students. The number of Kentucky students with limited English proficiency has grown from about 1,300 in 1990 to over 11,000 in 2005 (see Figure 2). This is an increase of over 700 percent during a period when overall enrollment was essentially flat. These students spread across most of Kentucky’s 176 school districts and speak over 100 languages. Spanish is the most frequently spoken language among these students (59 percent), followed by Japanese (5 percent), Bosnian (4 percent), Vietnamese (4 percent), Serbo-Croatian (4 percent), and Arabic (3 percent).(19)
Figure 2: Limited English Proficient (LEP) Students in Kentucky’s Public Schools
Schools do not ask students about their citizenship status because of a 1982 U.S. Supreme Court ruling, Plyler v. Doe, prohibiting school districts from refusing a public education to illegal aliens.(20) Consequently, there are no reliable data on the citizenship status of K-12 students; it must be estimated. Obviously, not all limited English proficient students are undocumented or illegal—our estimate is 15 percent.
Using national-level data from the Census Bureau, the Urban Institute has estimated that 40 percent of LEP students 5 to 19 years old are foreign-born.(21) In Kentucky, roughly 59 percent of LEP students are Spanish speakers,(22) and an estimated 43 percent of Hispanics are foreign-born.(23) If 43 percent of Hispanic and 40 percent of non-Hispanic LEP students are foreign-born, then about 42 percent of Kentucky’s LEP students are foreign-born ((0.59×0.43)+(0.41×0.40)=0.418 or 42 percent). The Pew Hispanic Center has analyzed the foreign-born population and estimates that Kentucky’s unauthorized population constitutes between 30 and 39 percent of the foreign-born.(24) To derive the estimated percentage of undocumented LEP students, we multiply the 42 percent of LEP students who are foreign-born times the midpoint estimate (35 percent) of foreign-born who are thought to be illegal and get 15 percent (0.42×0.35=0.147 or 15 percent).
Undocumented School-Age Children Enrolled in School. An estimated 97 percent of all Kentucky children between the ages of 6 and 17 are enrolled in school,(25) and national enrollment percentages for whites, blacks, and Hispanics are similar (see Table 1). Indeed, according to a 2004 U.S. Department of Education Report, “Language minorities enroll and are retained in elementary/secondary school at rates that are not measurably different from those of their counterparts who speak only English at home.”(26) However, the percentage of undocumented school-age children enrolled in school is likely lower than 97 percent because many Hispanic undocumented teens are in this country solely to work. One assessment is that among Hispanics aged 16 and older, the so-called emancipated youth, the percentage enrolled in school is probably less than 10 percent.(27) We use a weighted average to account for the portion of school-age children who are 16 or 17 years old, which is about 17 percent.(28) Assuming that 10 percent of 16- and 17-year-olds are enrolled in school and that the remaining children aged 6 to 15 attend at a rate similar to the total population (97 percent), it follows that 82 percent of undocumented school-age children are enrolled in school ((0.1×0.17)+(0.97×0.83)=.822 or 82 percent).
School-Age Children as a Percentage of the Undocumented Population. The Pew Hispanic Center estimates that unauthorized children constituted approximately 16 percent of all unauthorized individuals in 2005,(29) a slight increase from the 14 percent estimate for 2004.(30) However, this estimate includes all children under 18 years old, not just school-age children. Moreover, since unaccompanied adults typically account for the first wave of immigration into an area, there could be a higher proportion of illegal immigrant adults in Kentucky given the relative recentness of Hispanic immigration here. We estimate that undocumented school-age children constitute about 8 percent of the total undocumented population in Kentucky.
Using the average of the two estimates from the Pew Hispanic Center, we assume that unauthorized children constituted approximately 15 percent of all unauthorized individuals. Then we adjust the 15 percent estimate downward to exclude children not in school and get a revised estimate of about 9 percent (0.6×0.15=0.09 or 9 percent).(31) Moreover, to account for the likelihood of more unaccompanied adult immigrants in Kentucky compared with the U.S. average, we use the adult-to-child ratios from the 2000 Census as a proxy. According to the 2000 Census, around 35 percent of Hispanics nationally were under the age of 18, compared with about 31.5 percent in Kentucky. Since Kentucky’s percentage is 90 percent of the U.S. percentage (0.315÷0.35=0.9), we apply a similar factor to the 9 percent estimate of school-age children as a percentage of the undocumented population and arrive at about 8 percent (0.9×0.9=0.081 or 8 percent).
Our estimates of Kentucky’s undocumented population from 1990 to 2005 show that it is growing steadily and currently stands at about 26,000, just under the Pew Hispanic Center’s estimated range of 30,000 to 60,000. Moreover, with the exception of 2005, our estimates are quite close to the INS and Pew Hispanic Center estimates (see Figure 3). The notion that the undocumented population could be twice this value—as implied by the Bear Stearns report and other anecdotal estimates—is not supported by our data. However, the strength of this approach is its simplicity. Alternative data can be plugged into the formula to produce a different estimate. If one believes that different values should be used for any of the three factors, the method is sufficiently transparent that an alternative estimate can be easily generated.
Figure 3: Estimates of Kentucky’s Undocumented Population
Despite the simplicity and transparency of this approach, there are important caveats. First, the precision of our 2005 estimate belies the inherent uncertainty in estimating the size of the undocumented population. The estimates for the three antecedent factors are empirically based, but in some cases are educated guesses. We do not know, for example, how many emancipated youth attend school or whether the adult-to-child ratio from the 2000 Census is a good proxy for the unaccompanied adult-to-child ratio. Second, the estimated number of illegal immigrants can be quite sensitive to the values used for the three factors. For example, the estimated number of illegal immigrants in 2005 changes by 1,700, 300, or 3,700 by changing the percentage of LEP students who are undocumented, how many enroll in school, and their portion of the total undocumented population by 1 percentage point respectively.(32) The last factor, school-age children as a percentage of the total undocumented population, has the largest effect. By holding the other two factors at their likely values, 15 and 82 percent, but decreasing the percentage of school-age children constituting the total undocumented population from 8 to 7 percent, the estimated number of total undocumented increases by 3,700 from 25,600 to 29,300.
Despite these limitations, this approach imposes an internal consistency and level of rigor that most seat-of-the-pants estimates lack. With time, it is likely that better data on the undocumented population will be available, allowing us to refine our estimates. For now, however, this approach enables us to estimate the size of Kentucky’s undocumented population and to assess the plausibility of the antecedent factors used to estimate it. For policymakers and others attempting to respond to the presence of an undocumented population, this should bring some clarity and focus to an issue with contours that are otherwise amorphous.
* Mr. Childress is the Executive Director of the Kentucky Long-Term Policy Research Center. Return to text.
1 Ann Morse, Adam Blott, and Leya Speasmaker, “2006 State Legislation Related to Immigration: Enacted, Vetoed, and Pending Gubernatorial Action,” National Conference of State Legislatures, 7 June 2006, 19 June 2006 http://www.ncsl.org/programs/immig/06ImmigEnactedLegis.htm. Return to text.
2 U.S. General Accounting Office (GAO), Immigration Statistics: Information Gaps, Quality Issues Limit Utility of Federal Data to Policymakers, GAO/GGD-98-164 (Washington: GAO, 1998). Return to text.
3 Congressional Budget Office (CBO), Projections of Net Migration to the United States (Washington: CBO, 2006). Return to text.
4 We use the terms illegal, undocumented, and unauthorized interchangeably to describe this population. Return to text.
5 GAO, Illegal Alien Schoolchildren: Issues in Estimating State-By-State Costs, GAO-04-733 (GAO, 2004) and Undocumented Aliens: Questions Persist About Their Impact on Hospitals’ Uncompensated Care Costs, GAO-04-472 (GAO, 2004). Return to text.
6 CBO, “Cost Estimate,” S. 2611 Comprehensive Immigration Reform Act of 2006 (CBO, 2006). Return to text.
7 U.S. Commission on Immigration Reform, “Executive Summary,” U.S. Immigration Policy: Restoring Credibility, 1994, 27 June 2006 http://www.utexas.edu/lbj/uscir/exesum94.html. Return to text.
8 Carl Bialik, “Fuzzy Math on Illegal Immigration,” The Wall Street Journal 5 Apr. 2006, 8 May 2006 http://online.wsj.com. Return to text.
9 Michael Hoefer, Nancy Rytina, and Christopher Campbell, Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2005 (Washington: Department of Homeland Security, 2006) 21 Aug. 2006 http://www.uscis.gov/graphics/shared/statistics/publications/ILL_PE_2005.pdf. Also see Jeffrey S. Passel, Size and Characteristics of the Unauthorized Migrant Population in the U.S. (Washington: Pew Hispanic Center, 2006) 28 June 2006 http://pewhispanic.org/files/reports/61.pdf. Return to text.
10 Using a survey-based approach to estimate the 2000 Census undercount of Hispanics in Los Angeles County, Enrico Marcelli, currently on the faculty at the University of Massachusetts, Boston, found that all but about 10 percent were counted. The INS used his estimated 10 percent undercount when it estimated the number of unauthorized immigrants in the entire U.S. in 2003. Return to text.
11 Pew Hispanic Center, Estimates of the Unauthorized Migrant Population for States Based on the March 2005 CPS, Apr. 2006, 28 June 2006 http://pewhispanic.org/files/factsheets/17.pdf. The 2004 Kentucky estimate was between 20,000 and 35,000. Refer to Passel, Unauthorized Migrants: Numbers and Characteristics (Washington: Pew Hispanic Center, 2005,) 30 June 2006 http://pewhispanic.org/files/reports/46.pdf. Return to text.
12 The U.S. Immigration and Naturalization Service (INS) estimates for Kentucky in 1990, 1996, and 2000 are 4,000, 6,000, and 15,000, respectively. Refer to U.S. INS, Office of Policy and Planning, Estimates of the Unauthorized Immigrant Population Residing in the United States: 1990 to 2000 (Washington: INS, 2003) 30 June 2006 http://www.uscis.gov/graphics/shared/aboutus/statistics/Ill_Report_1211.pdf, and Legislative Research Commission (LRC), Immigration in Kentucky: a Preliminary Description, Research Report No. 305 (Frankfort: LRC, March 2002). Return to text.
13 Robert Justich and Betty Ng, “The Underground Labor Force Is Rising to the Surface,” Bear Stearns Asset Management Inc., 3 Jan. 2005, 28 June 2006 http://www.bearstearns.com/bscportal/pdfs/underground.pdf. Return to text.
14 Ron Crouch, Kentucky State Data Center, e-mail to the author, 26 July 2006. Return to text.
15 Lexington-Fayette Urban County Government (LFUCG), The Mayor’s Task Force for Hispanic Workforce Development, Findings & Recommendations (Lexington, Kentucky: LFUCG, 2005). Return to text.
16 Peter Laufer, “My New Kentucky Home,” Washington Monthly Jan.-Feb. 2005, June 2006 http://www.washingtonmonthly.com/features/2005/0501.laufer.html. Return to text.
17 Kentucky Department of Education (KDE), “Kentucky English Language Proficiency Standards,” KDE, Frankfort, 17 June 2006, 7 Aug. 2006 http://www.education.ky.gov/KDE/Instructional+Resources/High+School/Language+Learning/English+Language+Learning/Kentucky+English+Language+Proficiency+Standards.htm. Students who qualify for LEP services are defined as (1) individuals who were not born in the United States or whose native language is a language other than English; (2) individuals who come from environments where a language other than English is dominant; and (3) individuals who are American Indian/Alaska Native and who come from environments where a language other than English has had a significant impact on their level of English language proficiency; and who, by reason thereof, have sufficient difficulty speaking, reading, writing, or understanding the English language. Essentially, these are students whose command of English is below grade- or age-level peers. Return to text.
18 ([(11,200×0.15)/0.82]/0.08)=26,000. Return to text.
19 KDE, “LEP District Summary Home Language Survey 2003-2004,” LEP District Summary Reports (2003-2004) 8 Aug. 2006 http://www.education.ky.gov/. Return to text.
20 GAO, Illegal Alien Schoolchildren. Return to text.
21 Michael Fix and Jeffrey Passel, U.S. Immigration—Trends and Implications for U.S. Schools (Washington: Urban Institute, Jan. 2003) 8 Aug. 2006 http://www2.urban.org/UploadedPDF/410654_NABEPresentation.pdf. Return to text.
22 KDE, “LEP District Summary 2003-2004.” Return to text.
23 2000 Census data show 59,939 Hispanics in Kentucky with 25,605 being foreign-born (25,605÷59,939=0.427). Refer to Nolan Malone, Kaari F. Baluja, Joseph M. Costanzo, and Cynthia J. Davis, “The Foreign-Born Population: 2000,” Census 2000 Brief (Washington: U.S. Census Bureau, 2003) 24 Aug. 2006 http://www.census.gov/prod/2003pubs/c2kbr-34.pdf#search="foreign born population 2000". Return to text.
24 Passel, Unauthorized Migrants: Numbers and Characteristics, Background Briefing Prepared for Task Force on Immigration and America’s Future (Washington: Pew Hispanic Center, 2005) 8 Aug. 2006 http://pewhispanic.org/files/reports/46.pdf. Return to text.
25 We estimated this by using the Kentucky State Data Center’s population estimates by single year of age for 2004. The estimated number of children from 6 to 17 is 667,309. The Census Bureau’s American Community Survey shows Kentucky’s total school enrollment for grades 1-12 at 649,783 in 2004. Thus, the enrollment percentage is 97.4 percent (649,783÷667,309=.974). This finding is consistent with our analysis of 1998-2000 pooled Current Population Survey data, which shows that around 98 percent of non-Hispanics and 96 percent of Hispanics between 5 and 17 attend school, both nationally and in Kentucky. Return to text.
26 Steven Klein, Rosio Bugarin, Renee Beltranena, and Edith McArthur, Language Minorities and Their Educational and Labor Market Indicators—Recent Trends (Washington: National Center for Education Statistics, 2004) 10 Aug. 2006 http://nces.ed.gov/pubs2004/2004009.pdf. Return to text.
27 Ken Ison, “Federal Programs and Instructional Equity,” Kentucky Department of Education, e-mail to the author, 1 May 2006. Return to text.
28 This is the distribution for all children, not just Hispanic children. Return to text.
29 Passel, The Size and Characteristics of the Unauthorized Migrant Population in the U.S. (Washington: Pew Hispanic Center, 2006) 10 Aug. 2006 http://pewhispanic.org/files/reports/61.pdf. Return to text.
30 Passel, Unauthorized Migrants. Return to text.
31 Based on the 2005 American Community Survey, Hispanic children under age 5 (9,685) accounted for 41 percent of all Hispanic children under 18 (23,695). Thus, the school-age portion is about 60 percent. Return to text.
32 For example, the values used for the three factors are 15, 82, and 8, which yields an estimate of 25,600 undocumented. If we change the first factor from 15 to 16, but keep the other two factors at 82 and 8, the new estimate increases by 1,700 to 27,300. Return to text.