angle-up angle-right angle-down angle-left close user menu open menu closed search globe bars phone store

Too Big To Fail: Millennials on the Margins
Anita Sands and Madeline Goodman

Who Are Our Low-Skilled Millennials?

U.S. millennials have been labeled "the most ethnically diverse adult cohort in American history" and are projected to be the parents of an even more heterogeneous population over the coming decades.46 In Diversity Explosion: How New Racial Demographics Are Remaking America (2014), demographer William Frey chronicles how vital it is for us to take full notice of demographic trends affecting the millennial cohort. "It is critical to ask how well poised this increasingly younger, diverse population is—and will be—to sustain themselves and their communities," Frey warns. "They've got to be well trained. They've got to be ready to move into those high-tech jobs, the middle class jobs, or at least be able to support themselves and to support their communities."47 Despite the fears that the changing demographics at times spark in the short run, Frey argues, "investments in younger minorities—whom [some] may not yet see as their children and grandchildren—are crucial not only to the success of the nation's economy but also to future contributions to government programs like Social Security and Medicare." 48 In essence, then, the more we realize that the gaps in skills reflected both between and among groups are part of a larger problem of the whole people, the better off we will be.

Below we provide descriptive statistics from PIAAC for millennials by gender, race/ethnicity, and nativity/language status across the levels of proficiency discussed above. These data raise two related and important concerns. They reflect achievement gaps (particularly racial/ethnic gaps) that we—and many others—believe may ultimately stem from inequality in access to opportunity throughout the country.49 At the same time, in terms of absolute numbers, the problem of low skills among our millennial population is pervasive, with implications for all of us.


Overall, there is a slightly larger percentage of males than females in the millennial population (Table 3). The small difference in the average PIAAC literacy scores of millennial men (276) and women (278) was not statistically different. The relative parity of literacy scores for young adult men and women is in contrast to national and international assessments of school-age children that show a gender achievement gap in favor of girls.50 In PIAAC numeracy, however, male millennials scored higher than their female counterparts; also, a larger percentage of young adult women compared to men performed at or below Level 1 and at Level 2, and a smaller percentage performed at or above Level 3. This difference in performance between men and women in numeracy is mirrored in the performance of U.S. 15-year-olds internationally and 12th graders nationwide.51

The gender gap we observe in numeracy is likely correlated to structural and curricular issues in education, choice of major, occupational choices, and even work-related skill use, all of which warrant additional research.52

Table 3: Percentage distribution, average scores, and estimated number and percentage performing at select proficiency levels on PIAAC literacy and numeracy scales for millennials, by gender: 2012/2014


PIAAC data show that 59 percent of millennial respondents identified as White, 19 percent as Hispanic, 14 percent as Black, and 8 percent as Other race.53 Table 4 shows that Black millennials were approximately three times as likely as White millennials to perform at or below Level 1 in literacy, and Hispanic millennials were nearly four times as likely to do so. The 3-percentage-point difference between Blacks and Hispanics who performed at or below Level 1 in literacy was not statistically significant. At Level 2 in literacy, there was again no statistically significant difference in the performance of Blacks and Hispanics (46 and 40 percent, respectively), while White and Other race millennials had smaller percentages than both Blacks and Hispanics at this level. In numeracy, the pattern of performance between racial/ethnic groups differed somewhat. Although larger percentages of Whites and Other race millennials performed at or above Level 3 (52 and 47 percent, respectively) compared to Black and Hispanic millennials (and a smaller percentage at or below Level 1), there was no difference in the percentage between any of the racial/ethnic groups that performed at Level 2.

Table 4: Percentage distribution, average scores, and estimated number and percentage performing at select proficiency levels on PIAAC literacy and numeracy scales for millennials, by race and ethnicity: 2012/2014

The magnitude of the racial/ethnic gaps that we observe in the PIAAC data is deeply troubling. Studies indicate that the United States is increasingly segregated by income and has long been segregated by race/ethnicity, and that these overlapping structural realities likely play a critical role in both producing and perpetuating the achievement gaps observed across racial/ethnic groups and income groups at the K-12 level.54 These same inequalities may account for much of the skills gap seen in the PIAAC data. That these gaps are evident in the skills data of the young adult population should not in and of itself be surprising. What this communicates about inequality now and in the future, however, is of great concern.

Putting aside for a moment the underlying factors that perpetuate large racial/ethnic disparities in achievement in national and international assessments, the PIAAC data on young adult skills reveals that, in numerical terms, the problem of inadequate skills cuts across particular racial/ethnic groups.55 Over 16 million White millennials scored at or below Level 1 or at Level 2 in literacy, and almost 22 million did so in numeracy.

Nativity status and language

In order to account for the intersection of nativity status and language, our measure here uses a combined variable of native-born or native-language vs. foreign-born and foreign-language. This allows us to distinguish between those millennials for whom English is a first language and those for whom it is not. For the remainder of the report we will refer to the native-born or native-language as native-language and the foreign born and foreign language as foreign-language.

The growing heterogeneity of the U.S. young adult population is part of a global phenomenon that is touching a number of other countries, particularly many OECD countries. Though the United States has the largest absolute number of immigrants, foreign-language millennials make up only 9 percent of its overall millennial population. This is less than the percentage of foreign-language millennials in a number of other OECD countries, including, for example, Austria (13 percent), Canada (15 percent), Germany (11 percent), Norway (14 percent), and Sweden (15 percent) (Appendix Table A-4). Comparing the skills of native-language millennials to their counterparts in a subset of other countries with similar percentages of foreign-language millennials reveals several key points.56 First, in literacy, U.S. millennials who are native-language speakers scored higher on average than their native-language counterparts in only three countries (Ireland, Spain, and Italy). Second, in numeracy, U.S native-language speakers scored higher on average than their counterparts in only one country—Spain. Third, foreign-language millennials in the United States scored higher on average than their foreign-language millennials in three countries (Sweden, Spain, and Italy) in literacy and in two countries (Spain and France) in numeracy (Appendix Table A-5).

Across our subset of countries, native-language millennials scored higher on average than their foreign-language peers in both literacy and numeracy, and a smaller percentage of native language millennials performed at the lowest proficiency levels (Appendix Table A-6). This is largely an expected outcome given that PIAAC is administered across OECD countries in the native language of the host country. Comparative analyses have found that the gap between foreign- and native-language adults observed across all countries is related to country of origin, levels of educational attainment, age at immigration, and access to/experience with schooling in host countries.57 Over and above the issue of gaps in the performance of native language vs. foreign language, the number of U.S. native-language millennials in the low-skilled group is noticeably high. In literacy, nearly 8 million performed at or below Level 1, with an additional 23 million performing at Level 2. In numeracy, once again, these numbers are even larger. Nearly 16.4 million native-language millennials performed at or below Level 1, with an additional 24.4 million performing at Level 2 (Table 5).

Table 5: Percentage distribution, average scores, and estimated number and percentage performing at select proficiency levels on PIAAC literacy and numeracy scales for millennials, by nativity and language status: 2012/2014

We now turn from an examination of the demographic dimension of U.S. millennials and its relationship to skills to an investigation of how employment and participation in educational activities relate to skill levels.


46 William H. Frey, In the U.S., Diversity is the New Majority, Brookings Institution, March 6, 2015,

47 William H. Frey, interview by Tavis Smiley, PBS, February 17, 2015,

48 Frey, In the U.S.

49 Kirsch et al., Choosing Our Future; Sean F. Reardon, Demetra Kalogrides, and Ken Shores (2017), The Geography of Racial/Ethnic Test Score Gaps, Stanford University, Center for Education Policy Analysis,

50 F or results on NAEP reading by gender at grade 12, see; for international results on PISA reading literacy, see Some have posited that the lack of gender difference in PIAAC bears some relationship to the fact that literacy in PIAAC is defined differently than reading, or reading comprehension is in both NAEP and PISA. See Oddny Judith Solheim and Kjersti Lundetræ, "Can Test Construction Account for Varying Gender Differences in International Reading Achievement Tests of Children, Adolescents and Young Adults?—A Study Based on Nordic Results in PIRLS, PISA and PIAAC," Assessment in Education: Principles, Policy & Practice, (2016), doi:10.1080/0969594X.2016.1239612.

51 For results on NAEP mathematics by gender at grade 12, see For results on PISA mathematics, see

52 OECD Skills Outlook 2013; Danielle J. Lindemann, Gender and Numeracy Skill Use: Cross-National Revelations from PIAAC (Washington, DC: American Institutes for Research, 2015),; Fogg, Harrington, and Khatiwada, Human Capital Investments.

53 Race is a derived variable based on two questions on race and ethnicity in the PIAAC. We used the derived variable that includes four categories of race/ethnicity: White, Black, Hispanic and Other. A second derived variable breaks the data into a fifth category, Asian. However, due to reporting standards, using this version would reduce our ability to report data.

54 Ray Chetty, Michael Stepner, Sarah Abraham, Shelby Lin, Benjamin Scuderi, Nicholas Turner, Augustin Bergeron, and David Cutler, "The Association Between Income and Life Expectancy in the United States, 2001-2014," Journal of the American Medical Association 315, no. 16, (2016): 1750-1766, doi:10.1001/jama.2016.4226; Fabian T. Pfeffer and Alexandra Killewald, Generations of Advantage. Mulitgenerational Correlations in Wealth from a Life-Course Perspective (2017),; Martin Carnoy and Emma Garcia, Five Key Trends in U.S. Student Performance (Washington, DC: Economic Policy Institute, 2017), retrieved from; Raj Chetty, Nathaniel Hendren, and Lawrence F. Katz, "The Effects of Exposure to Better Neighborhoods on Children: New Evidence; from the Moving to Opportunity Experiment," American Economic Review 106, no. 4 (2016): 855-902; Reardon, Kalogrides, and Shore, Racial/Ethnic Test Score Gaps; Douglas S. Massey and Jonathan Tannen, "Segregation, Race, and the Social Worlds of Rich and Poor," in The Dynamics of Opportunity in America: Evidence and Perspectives, ed. Irwin Kirsch and Henry Braun (New York: Springer, 2016), 13-33; Gary Orfield, John Kucsera, and Genevieve Siegel-Hawley, E Pluribus ... Separation: Deepening Double Segregation for More Students, UCLA, The Civil Rights Project,

55 Sean F. Reardon, "The Widening Academic Achievement Gap between the Rich and the Poor: New Evidence and Possible Explanations," in Whither Opportunity? Rising Inequality, Schools, and Children's Life Chance (New York: Russell Sage Foundation, 2011), 91-116.

56 For comparative purpose, we look here at the subset of OECD countries that have equal or greater percentages of foreign-born/foreign-language millennials. These include the following: Australia, Austria, Canada, Denmark, France, Germany, Ireland, Italy, Netherlands, Norway, Spain, Sweden and England/Northern Ireland. For data on native language/foreign language millennials across these countries, see Appendix tables A-4, A-5, and A-6.

57 Goodman, Sands, and Coley, America's Skills Challenge; Jeanne Batalova and Michael Fix, Through an Immigrant Lens: PIAAC Assessment of the Competencies of Adults in the United States (Washington, DC: Migration Policy Institute, 2015),; Jeanne Batalova and Michael Fix, Literacy and Numeracy Skills of Second-Generation Adults: A Comparative Study of Canada, France, Germany the United Kingdom and the United States (Washington, DC: Migration Policy Institute, 2016),; Mark Levels, Japp Dronkers, and Christopher Jencks, "Contextual Explanations for Numeracy and Literacy Skill Disparities between Native and Foreign-Born Adults in Western Countries," PLOS ONE,*.0172087; John Sabatini, Understanding the Basic Reading Skills of U.S. Adults: Reading Components in the PIAAC Literacy Survey (Princeton, NJ: Educational Testing Service, 2015),