|Trends and Challenges for Work in the 21st Century|
Opinions and views in these papers are those expressed by the author(s). They are not to be taken as expressions of support for particular positions by the Department of Labor. Please do not cite these papers without prior permission of the author(s).
Functional Literacy and Labor Market Outcomes
Robert I. Lerman and Stefanie R. Schmidt
The Urban Institute
100 M Street, NW
Washington, DC 20037
This paper was delivered at an Urban Institute conference sponsored by the US Department of Labor, Office of the Assistant Secretary for Policy, "Helping Low-Wage Workers: Policies for the Future," May 6-7, 1999. The authors thank Edgar Lee for his solid research assistance. The opinions expressed in this paper are not necessarily those of the Urban Institute or its sponsors.
In todays strong job market, the returns to skill are rising quickly. More employers are requesting that workers master basic reading, writing, and mathematical skills. This paper uses results of the National Adult Literacy Survey to examine the connection between these types of "functional literacy" skills and weak job market outcomes. We find that a substantial share of the US population lacks the basic skills needed for a wide range of middle to high-wage jobs. Workers with limited functional literacy account for a disproportionate share of low-wage, prime-age workers. Even among workers with the same characteristics in terms of education, race, marital status, age, health status, and region of the country, those with low literacy levels fall substantially behind in the labor market than those with adequate or high literacy levels. Still, the connection between low wages and low skill is far from complete; many low wage workers (40 percent of males and 50 percent of females) have average or above-average levels of functional literacy.
This paper uses results of the National Adult Literacy Survey to examine the connection between "functional literacy"defined by basic reading, writing, and mathematical skillsand weak job market outcomes. We not only find that a substantial share of the US population lacks the basic skills needed for a wide range of middle to high-wage jobs, but we also determine that low literacy levels substantially weaken labor market outcomes, especially among women.
Keywords: Literacy, Skills, Low-wage Workers, National Adult Literacy Survey, Job Market
Adequate basic skill is increasingly viewed as necessary for workers to obtain a good-paying job. More jobs are requiring skills that call for mastery of basic reading, writing, math, and work place skills (Levy and Murnane, 1998). New technologies are eliminating less-skilled positions and expanding demand for people who can work with computers and deal with an increasingly complex service sector. These technical changes have no doubt played a major role in the rising return to skill and the decline in real wages for less-skilled American workers.
While the country has embraced the strategy of raising skills as a primary method for raising the capacity of workers to earn better wages, the specific shortcomings of the work force and their connection to the low wage problem are not well documented. Most studies rely on information about years of schooling, not on data documenting reading, writing, and math competencies. Certainly, years of school completed are positively correlated with skill and wages, but educational attainment fails to capture the wide variability in competencies among those with the same level of education. Evidence from analyses of wage inequality shows that the variability in wages within a single educational category is almost as high as the overall level of wage variability (Lerman, 1998).
If educational attainment is a highly imperfect indicator of low wages, what about direct measures of skill? Perhaps some of the workers earning low wages in spite of having completed high school simply lack basic skills. On the other hand, the basic skills of many workers without a high school degree may be high enough to give them access to well-paying jobs.
The distinction between skill and education may explain part of the wage erosion among workers with low levels of formal education. Older cohorts of the population were much less likely to have graduated high school than younger cohorts. Of those born in the years 1938 to 1943, only about 76 percent graduated high school; of those born between 1973 and 1978, 88 percent had graduated high school. Put another way, dropouts in an earlier period were a much more common phenomenon; the typical dropout was at about the 12th percentile of the educational distribution. Today, dropouts are only at the 6th percentile. If literacy level is an independent dimension of skill, then todays dropouts, who represent more of the bottom tail of the distribution, might exhibit less skill than comparable dropouts 30 years ago. One study found that taking account of the changing rankings of those without a college education explains some of the deterioration of earnings of workers with less formal education (Rosenbaum, 1998). Weak English language skills among immigrants is another reason those with apparently adequate educational levels may lack basic skills.
Given the incomplete connection between measures of education and skill, policymakers attempting to target training resources on those lacking basic skills require more reliable indicators than educational attainment alone. This paper examines the levels of basic skills directly and how these skills overlap with low wages. Using data from the National Adult Literacy Survey (NALS), we focus on three major questions:
- Do most low-wage workers lack adequate levels of functional literacy? What are the basic skills of low-wage workers? Do adequate basic skills allow workers with low formal education to earn medium or high wages? How do weak basic skills overlap with low weekly earnings, high unemployment, and low amounts of work over the year?
- How have the basic skills of high school graduates and high school dropouts changed across birth cohorts? As the population raised its educational attainment, have individuals lacking any college education become more concentrated in the bottom of the distribution of abilities?
- How much of the variation in wages occurs within groups of the same literacy level? Do literacy categories have more salience for determining low wages than educational attainment levels? What is the impact of very limited literacy on the probability of having a very low wage job?
An expanded human capital strategy is central to reversing the decline in wages for less skilled Americans. The rising return to education strongly suggests the importance of skill to employers. The increase in the relative cost of hiring well-educated workers over less-educated workers is not deterring employers from willingly paying the premium to add workers with high skills.
In 1992, the National Adult Literacy Survey was administered to document the literacy levels of the adult population of the United States. The survey was funded by the National Center for Education Statistics at the US Department of Education, and was administered by the Educational Testing Service in collaboration with Westat, Inc. The survey was administered in-person to a nationally representative sample of approximately 25,000 individuals ages 16 or older who spoke either English or Spanish. The survey included two sections. The first section was a background questionnaire that gathered information on demographic characteristics, employment, educational background, immigrant status, and the use of reading and math skills in work or personal life.
The second part of the survey was a test designed to measure respondents ability to apply math and reading skills to real-life tasks. Only individuals who could read English took the literacy test. The literacy tests were not identical. Rather, the respondents were given one of 26 possible literacy tests. Each of the 26 tests consisted of 35 and 41 tasks that were drawn from a set of 166 possible tasks in an assessment pool, and each test took approximately 15 minutes to complete. The test administration design (known as standard matrix sampling) was chosen so that the survey could characterize the literacy level of the overall population or population subgroups in a minimal amount of time.
Each individuals literacy score is generated from a statistical model that predicts literacy score from performance on the subtasks. Based on the model, each individual received a score on the NALS from 1 to 5, 1 being the lowest level of literacy, 5 being the highest. Individuals received an overall functional literacy score, but also received a subscore in each of three areas: prose (reading), document (ability to read charts and graphs), and quantitative (the ability to perform a mathematical operation in a real world context).
It is easiest to understand the overall functional literacy score in terms of what types of questions a person at that level of literacy can answer. Most individuals at the lowest level of literacy, level 1, are able to do very simple tasks such as locate the expiration date on a drivers license, total a bank deposit slip, or sign their names. But they are unable to do level 2 tasks, which include locating an intersection on a street map, understanding an appliance warranty, or totaling the costs from an order. Individuals at literacy level 2 can perform level 2 tasks, but cannot perform level 3 tasks such as writing a letter explaining an error on a credit card bill, using a bus schedule, or using a calculator to determine a 10 percent discount.
The analysis in this study concentrates on the groups with only level 1 or level 2 literacy. Although the tests measure prose, document, and quantitative aspects of literacy separately, the correlation between types of tests is very high, over .9. Thus, we focus on a composite measure drawn from all three aspects of literacy. We group individuals into composite literacy levels 1 through 5.
The distribution of the population across literacy categories in Table 1 reveals that over one-third (37 percent) of the 25 to 54 year-old US population did not reach above literacy level 2. About 14 percent attained the lowest functional literacy. The distributions of functional literacy were similar among men and women, though men performed somewhat better than women on the NALS tests. About 29 percent of men but only 23 percent of women demonstrated a literacy level of 4 or 5.
Not surprisingly, levels of functional literacy matched closely with levels of educational attainment. The overwhelming majority (83 percent) of workers lacking a high school degree or GED equivalent fell into the bottom two literacy categories compared to only 6 percent of college graduates. It is disappointing that over 40 percent of those at middle levels of education (those with a high school diploma or GED) could not score beyond the level 2 threshold. Another disappointment is that about 20 percent of those with some years of college but no BA displayed no more than level 2 literacy.
While functional literacy patterns clearly increase with education, educational groups overlap a good deal with respect to functional literacy. For example, about 57 percent of high school dropouts had literacy levels at or above the literacy levels of the bottom 40 percent of high school graduates. The top 58 percent of high school graduates reached level 3 or 4levels attained by the top 77 percent of those with some college. And 35 percent of college graduates scored no higher (at level 3 or below) than the top 58 percent of high school graduates.
In short, the literacy tests have a close connection with educational levels, but the relationship is not one-to-one. While only a tiny proportion of high school dropouts reach the literacy levels common among those with some college, the overlap in functional literacy is substantial between high school graduates and those with some amount of college but not the BA.
The NALS includes data on annual and weekly earnings, but do not provide weekly hours information for all workers. As a result, we chose to define low earnings in terms of earnings in the prior week. Low-wage workers were those with earnings more than $20 per week but less than $218 per week (or less than about $6.20 per hour at 35 hours per week); medium-wage workers earned between $218 and $704 per week (about $20.00 per hour at 35 hours per week); and high-wage workers earned more than $704 per week.
The employment status categoriesemployed, unemployed, or not in the labor forcerefer to an individuals status as of the prior week, as does the official measure of employment and unemployment.
Connections Between Functional Literacy, Employment, and Low Earnings
Most of the research evidence concerning the importance of skill in the job market comes from analyses of educational attainment and years of potential work experience. In some cases, analysts have been able to incorporate estimates based on actual work experience, actual training, and educational field as well as years of schooling and degree status. The studies that have examined the relationship between labor market outcomes and direct measures of skill or academic competency (through performance on tests) typically rely on data drawn from only a young cohort of the populationthose ages 14 to 21 in 1979 (31 to 38 in 1996).
This analysis draws on the NALS to investigate the relationship between low wages and low levels of functional literacy for the entire adult population. We consider the interaction between unemployment and functional literacy as well.
Concentration of Low Wage Workers by Literacy and Education
The first issue is whether the low-wage worker group is primarily made up of people with minimal functional competency. We focus on the prime-age adult labor force, or workers ages 25 to 54, in order to abstract from low wages commonly earned by students, youth, and some part-time elderly workers. To what extent are prime-age, low-wage workers concentrated in the low literacy categories? Table 2 answers this question, showing that 21 percent of low-wage workers were at the bottom literacy level and 28 percent were at the second lowest level. By comparison, 35 percent of the total prime age labor force fell into the bottom two categories.
The connection between low wages and low functional literacy is much closer among men than among women. About 60 percent of low-wage male workers but only 45 percent of low-wage female workers fell into the bottom two literacy levels. In part, this differential comes about because we use the same low wage threshold for men and women. Since women earn lower wage rates than men do, a fixed wage threshold will incorporate a higher share of female than male workers. Given that only the bottom 9 percent of the male distribution but 24 percent of the female distribution fell below our low-wage threshold, it is not surprising that low-wage male workers are more concentrated at the bottom of the functional literacy distribution as well.
When we examine all low-wage workers, we find slightly lower concentrations in the lowest literacy groups than among prime-age, low-wage workers. About 45 percent of low-wage workers of all ages fell into the bottom two literacy categories, as compared to 37 percent of the labor force, ages 16 and over. The comparable figures for prime-age workers are 49 percent of low-wage workers and 35 percent of all workers. In part, this differential reflects the fact that while older groups have lower than average literacy levels, they have more work experience that allows them to move out of the low-wage segment of the market.
Unemployment is more closely linked to low literacy levels than are low wages. About 60 percent of the nations unemployed in 1991 fell into the lowest two literacy categories, as compared with 35 percent of the employed and 45 percent of low-wage workers.
Low-wage workers are at least as concentrated in the low education categories as they are in the low literacy categories. Among high school dropoutswho make up the bottom 14 percent of workersthe proportion of low-wage workers was 1.57 times the proportion of all workers. For those in the lowest literacy category (about 12 percent of workers), the ratio of low-wage to all employed workers was virtually identical (1.59).
Most but not all of the low-wage workers in low literacy categories have low levels of formal education. Nearly two-thirds of low-wage workers at level 1 literacy were high school dropouts. However, one in three had completed high school or even some college. At the next literacy category (level 2), about half of the low-wage workers had earned a high school diploma or GED and another 23 percent had completed at least some college. About 38 percent of low-wage workers both had low functional literacy and no more than a high school degree. By implication, over 60 percent of workers had either completed some college or were in the top 26 percent in terms of functional literacy.
Incidence of Low Wages and Unemployment by Literacy and Education Groups
Having asked about the share of low-wage workers with low literacy skills, we now turn to the question: what share of less skilled workers experience low wages or unemployment? Each approach yields distinct outcomes for targeting. If low-wage workers are highly concentrated in low literacy categories, then targeting on low literacy groups will allow policies to reach most low-wage workers. However, whether or not the benefits from targeting low literacy groups reach mostly low-wage workers depends on the proportion of low literacy groups with low wages and high unemployment. Alternatively, targeting on low-wage workers will channel benefits primarily to low literacy groups if low-wage workers are mostly those with low skills. But, low-wage targeting may not reach most low literacy workers unless the incidence of low wages in low literacy groups is high.
The evidence reveals sharp differences by gender in the proportion of low literacy workers with low wages. As Table 3 shows, 40 percent of workers in the lowest literacy category earn low wages. The gender gap in this proportion is quite large; of those employed at the lowest literacy level, 56 percent of employed women and 30 percent of employed men had low wages. The differential is nearly as high among those at level 2 in functional literacy. About 41 percent of women but only 17 percent of men at this functional category experienced low wages.
Attaining functional literacy at level 3 or higher substantially reduces but does not eliminate the chances of earning a low weekly salary. While only 10 percent of prime-age men with moderate to high literacy levels earned low wages, the figure for women was 28 percent. Most of these women earned low weekly wages because they worked part-time. Of the women at literacy levels 3 or higher and working full-time, only 12 percent earned low wages.
Unemployment rates varied significantly by literacy level but less so by gender. Note in Table 3 that the unemployment rate among women at the lowest literacy level was 20.5 percent, or over three times the 6.2 percent rate experienced by those at literacy level 3. At literacy level 1, males faced a 15.5 percent unemployment rate, about 5 points below the rate for females. However, at all other literacy levels, the male unemployment rate exceeded the female unemployment rate.
The gaps between high and low literacy levels in the employed share of the population were even higher than the unemployment differences. Again, the differentials were higher among women than among men. The difference in employment between those at level 3 and those at level 1 was about 15 percentage points for men and 30 percentage points for women.
The interaction between wages, incomes, and literacy is complex. First, it is noteworthy that 9 percent of employed prime-age women and 6 percent of men worked at low wages and lived in families with incomes below 150 percent of the poverty line. Among 25 to 54 year-old workers, the proportions were even lower. Although women were more likely than men to earn low wages, the proportions of men and women with both low wages and low incomes were similar.
Second, the incidence of low wages and low incomes was several times higher among those at the lowest functional literacy level (level 1) than among those with at middle levels or higher (level 3 or higher). Among the employed at level 1, 22 percent of women and 14 percent of men earned low wages and lived in low income families. For workers at level 3 or above, the incidence of low wages and low incomes was only 3.5 percent for men and 6.5 percent for women. Third, about 70 percent of workers at the lowest literacy level lived in families with incomes above 150 percent of the poverty level. On the other hand, literacy levels were highly associated with low income. Among all prime-age individuals, the proportion living in low income families declines from 35 percent among the least literate group to 6 percent for the group with above average literacy.
Cohort-Age Patterns of Functional Literacy and Educational Attainment
Employers increasingly are expressing concern over the qualifications of young workers entering the job market. The 1983 report, A Nation At Risk, was one of many publications that dramatized the widening gap between job requirements and the skills of the work force. Even though educational attainment has increased rapidly over the last few decades, achievement tests for entire cohorts reveal little progress (Murnane and Levy, 1996). Thus, at any given level of schooling, todays students may know less than earlier cohorts.
One way of assessing this view is to examine the functional literacy levels of various age groups at each level of education. The math, reading, and comprehension capacities might have worsened within each educational category either because the quality of teaching and schooling has declined or because the average ability level of those in a particular educational category has declined. If innate ability and schooling are strongly and positively correlated so that workers at the top 20 percent in education are likely to be near the top 20 percent in ability, then changes in the educational distribution will alter the average percentile of ability of people in that category. Suppose that high school graduates represented the 30th to the 60th percentiles of the educational distribution of 55 to 64 year-olds and the 10th to the 40th percentiles of the educational distribution of 25 to 34 year-olds. In this case, the average ranking of more recent high school graduates would be at the 25th percentile, well below the 45th percentile average ranking for older high school graduates. Given their lower ranking, we might expect young high school graduates to have weaker literacy skills than older high school graduates.
Surprisingly, the NALS data reveal that 25 to 34 year-old high school graduates are achieving higher functional literacy levels than those achieved by 55 to 64 year-old high school graduates. About 15 percent of older graduates had only level 1 literacy and another 40 percent reached only level 2, as compared to 12 percent of younger graduates at level 1 and 32 percent at level 2. Put another way, 56 percent of 25 to 34 year-old graduates but only 45 percent of 55 to 64 year-old graduates operated at literacy level 3 or higher. The advantage for young graduates is unexpected, given the differential in percentile rankings of high school graduates in the two groups. The average percentile of high school graduates among 25 to 34 year-olds was 23 percent, only modestly lower than the 29th percentile ranking for 54 to 65 year-old high school graduates. Comparing 25 to 34 year-olds with 45 to 54 year-olds reveals a slight advantage for the older group of high school graduates, but the differential is small. Overall, since the data show that more recent high school graduates have literacy skills as high as high school graduates of an older vintage, the evidence indicates that the abilities conferred by a high school diploma have not deteriorated. However, because employers are increasingly demanding high skills, the abilities of todays high school graduates may be less well-suited to the job market than previous cohorts of graduates.
Functional Literacy, Education, and Other Factors Affecting Low Wages
Literacy and education are often connected with other disadvantages that lead to low wages and unemployment. Growing up in a poor neighborhood, lacking access to job networks, having little work experience, being in poor health, and living with minority, immigrant, or unmarried status may all contribute to weak labor market outcomes. This section distinguishes between the roles of literacy and educational attainment as distinct from the effects of several other factors that also lead to low wages.
We specify sets of probit equations estimating the determinants of employment and low wages (given employment) for males and female workers. Each equation includes age, race, birth place (foreign or US), region, marital status, health status, and a measure of education or literacy. For each dependent variable and gender, we specify educational capabilities in one of three ways: 1) level of functional literacy;
2) educational attainment; or 3) both literacy and educational level. Tables 4, Table 5, Table 6 and Table 7 present the impact estimates of literacy, education, and other personal characteristics on employment and low wage status of 25 to 54 year-olds.
The results reveal the significance of functional literacy in labor market outcomes as well as the strikingly higher impacts of literacy and education among women than among men. Even after taking account of differences in health status, marital status, age, foreign birth, and race or Hispanic origin, workers with low functional literacy are less likely to become employed and, if employed, more likely to find themselves earning low wages.
We begin by looking at men. According to Table 4, falling into either of the two lowest functional literacy categories reduces the probability of holding a job by about 5 percentage points. This reduction is similar in size to the effect of being a high school dropout or earning a GED but no high school diploma. Given that the employed share of men in this age group in 1992 was about 87 percent, a 5 percentage point reduction in job-holding would be enormously significant. Surprisingly, the negative effect of low functional literacy on male employment is slightly higher among those at literacy level 2 than at literacy level 1. The reason may be that level 1 literacy individuals are more willing to work at low wages than are individuals at literacy level 2.
For men, when functional literacy and educational attainment appear in the same equation, the impact of literacy on male employment falls, but remains negative and
significant at level 2. The effect of educational attainment among those with the same literacy appears somewhat stronger than the effect of functional literacy among those with the same educational attainment. One possible explanation is that literacy tests do not effectively capture differences in academic competencies learned in school. Another possibility is that individuals who complete more schooling are more reliable and more motivated workers, traits highly valued by employers.
Functional literacy has a larger impact on the incidence of low male wages than on male employment. Note in Table 6 that falling into the lowest literacy category raises the proportion of workers in low wage jobs by 12.5 percentage points relative to level 3. Male workers in level 2 showed a 5 percentage point higher incidence of low wages than those reaching level 3. These effects are large relative to the overall 11 percent share of male workers earning low wages. Low educational attainment raises the likelihood of low wages about as much as does low functional literacy. Including both literacy and education measures reveals a significant role for each. Holding constant educational levels, the effect of being at level 1 instead of level 3 literacy is to raise the incidence of low wages by 7 percentage points.
Taking account of education and functional literacy sharply reduces the measured impact of race on employment and low wage status. The size of the measured employment and wage disadvantages that can be directly attributed to race and Hispanic depends a great deal on the inclusion of education and literacy indicators. When we control only for age, region, marital status, health status, and immigrant status, the effect of the variable representing an individual status as black reduces employment rates by 6.9 percentage points and raises the incidence of low wages (among the employed) by 10 percentage points. After controlling for education but not literacy differences across individuals, the effects of race falls to a 4.8 point reduction in employment and 3.7 point increase in low wage status. Adding the functional literacy variables lowers the employment effect by a further one-third (to 3.2 points) and diminishes the wage effect to by over two-thirds (to 1.1 points). Similar declines occur for the effects of Hispanic status on employment and wages, as the estimating equations include education and then functional literacy. The incidence of low wages rises by 3.2 percentage points among Hispanic men when educational attainment is the only measure of academic competency. Taking account of functional literacy moves the Hispanic effect to a small level that is not statistically significant.
The role of functional literacy is substantially larger among women than among men. Note in Tables 5 and 7 that, holding other factors equal, falling into one of the lowest two functional literacy categories exerts a far greater impact than does being in the lowest two educational attainment groups. Given an individuals educational attainment as well as her age, race, region, immigrant status, health status, and marital status, the effect of level 1 literacy (relative to level 3) is to reduce employment probabilities by 10 percentage points and to raise the incidence of low wage by nearly 30 points. For the level 2 literacy category compared to level 3, the effects were a 6 percentage point reduction in employment and a 10 percentage point increase in low wages (among the employed).
Again, taking account of functional literacy made the labor market returns for black and Hispanic women look more favorable than they appear on the basis of education alone. Adding the functional literacy variable causes the employment disadvantage for black women of 4.5 percentage points become an effect that is not statistically significant. Also, the variable representing an individuals black status lowers the incidence of low wage by 8.7 percentage points when education is included but by 14.5 percentage points with the addition of functional literacy.
While functional literacy and education play critical roles in determining labor market outcomes, the earnings distributions across categories do overlap. For example, the top 25 percent of male earners at functional literacy level 2 make more money than the bottom 50 percent of earners at functional level 3. Similarly, the top 25 percent of literacy 1 earners received more than the bottom 50 percent of workers at level 2. At the same time, only 10 percent of workers at functional literacy level 1 earned as much as the bottom 50 percent at higher literacy levels.
The overlap in earnings distributions occurs among women workers as well. The top third of women in literacy level 2 earned as much as top half of women workers at level 3. Even the top 20 percent of workers at the functional literacy 1 level earned as much as the bottom 40 percent of workers at higher levels of literacy.
A substantial share of the adult US population (25-54 year-olds) lacks the basic skills expected for a wide range of well-paying jobs. About 14 percent cannot systematically perform such tasks as understanding an appliance warranty or totaling a list of costs. Over one-third (37 percent) of the US population cannot perform such activities as writing a letter explaining an error on a credit card bill, using a bus schedule, or using a calculator to determine a 10 percent discount. As this paper documents, these skill shortfalls are closely linked to weak job market outcomes.
Low wage jobs are concentrated among young workers and those with low levels of education and/or functional literacy. Although employed men under age 25 accounted for about 20 percent of all male workers, they represented 54 percent of the men with low earnings. Of male workers between the ages of 25 and 54, those at the lowest two literacy categories made up about 30 percent of the workers but 60 percent of the prime-age male workers earning low wages. Low wages are concentrated among women workers who are young or have low functional literacy, but the extent of concentration is lower than among men.
While workers with limited literacy skills account for a disproportionate share of low wage, prime-age workers, many of the low wage workers have average or above average levels of functional literacy. Even in the case of men, about 40 percent of low-wage workers in the 25 to 54 age group are workers with literacy level 3 or higher. The figure for women is over 50 percent.
The incidence of low wage work or non-employment is quite high among the least literate, but a large share of such workers manage to do better. At the lowest literacy level, about 30 percent of male workers and 56 percent of female workers had low earning positions as of 1992. Conversely, most men and nearly half the women at the lowest literacy levels managed to escape falling into the low wage category.
Low literacy levels clearly and substantially weaken labor market outcomes, especially among women. While tests measuring prose, document, and quantitative capabilities do not fully capture skill levels related to jobs, they exert independent impacts on employment and wages, over and above the effects of schooling and personal characteristics.
This papers evidence on the connection between functional literacy and job market outcomes has concrete implications for policy. First, raising levels of literacy of workers will almost certainly improve their prospects in the labor market. Although existing adult literacy programs have at best a mixed record of success, the main reason is their inability to retain participants long enough to raise skills, not the inability of workers to translate higher skills into improved job market outcomes. Second, targeting strategies may vary by gender. Since most low-wage adult male workers lack solid literacy skills, targeting interventions on groups in low-wage occupations or industries may be most appropriate. On the other hand, the majority of adult men with low literacy skills earn moderate or high wages. In the case of women, targeting on low skills will usually mean reaching low-wage workers. Over half of the lowest literacy group and 41 percent of the next lowest literacy group earn low wages or cannot find employment at all. A third implication is that literacy skills and education appear more important than race or Hispanic origin in predicting weak job market outcomes. Although anti-discrimination policies remain important for individual justice and perhaps for achieving gains at the middle and high levels of the market, skills are critical for helping minorities obtain a job and move beyond low-wage work.
Committee for Economic Development. The Employers Role in Linking School and Work. Committee for Economic Development. New York. 1998.
Levy, Frank and Richard Murnane. Teaching the New Basic Skills : Principles for Educating Children to Thrive.