• Number of Jobs Held in the Early Career
• Timing of the Transition to Stable Employment
• The Dynamic Perspective of the School-to-Work Transition

The perception that non-college-bound youth mill about in the labor market in the early years after leaving school is a statement about the dynamics of employment. The early career is characterized as a period when youth hold numerous jobs, many for short periods of time. Thus, we need to be able to measure the number of jobs held in the early career and account for the time spent in any given job. These inherently dynamic measurements are not captured in the snapshot view of activity status of Section 3, which considers only the fraction working or in school at any point in time. A major advantage of the NLS-Y data is that they enable us to go beyond such a static view to consider the school-to-work transition in this dynamic framework.

In this section, we use the information in the NLS-Y employment histories to further evaluate the employment experiences of young men and women in their early labor market career. We first examine the transition process in terms of the distribution of the number of jobs held at successive ages. We then consider the transition to stable employment, which is measured by job duration. We continue to analyze each SLG separately, first examining the patterns for men and then examining the patterns for women.

Number of Jobs Held in the Early Career

Several recent studies demonstrate that young adults hold a surprisingly large number of jobs early in their labor market career (Bureau of Labor Statistics [BLS], 1992, 1993; Topel and Ward, 1992; Veum and Weiss, 1993). For example, tabulations from the NLS-Y by the Bureau of Labor Statistics reveal that young men hold an average of 7.6 jobs between their 18th and 30th birthdays; the count is only slightly lower (7.3 jobs) for women (BLS, 1993). In this subsection, we examine this issue by considering the mean number of jobs held by youth at each age, separately for the four SLGs. Since the mean obscures the spread in the distribution of jobs held across youth, we also report the number of jobs held by the median youth, as well as by the young person at the 25th and 75th percentiles of the jobs-held distribution. Thus, we can assess the experience of the "typical" youth, and how different his or her experience is compared with a counterpart who is at the bottom or top of the jobs-held distribution. In our analysis, jobs are counted only after school leaving and are defined as employment with a given employer.[15]

Patterns for Men

For each SLG and age, Table 4.1 presents the mean number of jobs held by young men, as well as the number of jobs held at the 25th, 50th, and 75th percentiles of the distribution of the number of jobs held. Table 4.1 confirms that young men hold a large number of jobs in the years immediately after leaving school, a result that is consistent with the findings reported elsewhere (BLS, 1992; Topel and Ward, 1992; Veum and Weiss, 1993). For example, by age 28, high school dropouts have held an average of 8.6 jobs since they left school. The average number of jobs held at the same age declines for each successive SLG, with college graduates having held 3.6 jobs by age 28.

There are sharp differences by SLG in the number of jobs held during the early career.[16] Most notably, the number of jobs held by high school dropouts is larger than for the other SLGs. The median male in this group has held six jobs by age 24 and eight jobs by age 28. A high school dropout at the 75th percentile of the jobs-held distribution, in contrast, has held nine jobs by age 24 and more than ten jobs by age 28. As a summary measure, these figures represent about one job every other year at the median and one job per year at the 75th percentile.

The high school graduate group starts working about one year later (on average) and holds fewer jobs. After a year, it is accumulating about half a job per year at the median, less than one-third of a job per year at the 25th percentile, and about one job per year at the 75th percentile. Similar patterns exist for the some college group. Finally, the college graduates start about four years later than the high school graduates, and they accumulate new jobs the most slowly. Exact comparisons are difficult because most of them were still in their second or third job by the end of the survey.

Differences by Race/Ethnicity

Table 4.2 repeats the data on number of jobs held just illustrated for all men, separately by the three race/ethnic groups considered previously. Within the three groups of men, it is still the case that both the mean and median number of jobs held is higher at each age for lower SLGs. At the same time, for Hispanic and white males the contrast between high school dropouts and high school graduates is sharper than it is for black males. For example, at age 26, the average black male high school dropout has held about three-fourths of a job more than the average high school graduate, whereas the average white or Hispanic dropout has held almost 2 more jobs than his counterpart who graduated from high school. These differences may result from differences in the age of school leaving across the three groups, or to differences in job turnover within groups.

The more interesting contrast looks at differences within each SLG by race/ethnicity. These differences are greatest for high school dropouts: White males on average have held more jobs at each age, and black males have held the fewest jobs. For example, at age 27, white male dropouts have held an average of 9.1 jobs, compared with 8.4 jobs for Hispanic males and 7.0 jobs for black males. These differences persist, as well, at the bottom and top of the jobs-held distribution. For the two other SLGs, blacks continue to accumulate fewer jobs at each age, while the distribution of jobs held for Hispanic and white males look virtually identical through the early-to-mid-twenties.[17]

How should we interpret these differences by race/ethnicity? Do these numbers indicate that whites mill about more in the labor market in their early career, while blacks have the most stable employment experiences? By viewing these numbers in light of the static analysis in Section 3, it is possible to conclude that black men hold fewer jobs not because they hold each job for a longer period, but because they spend more time unemployed or out of the labor force.[18] This issue will be explored further when we examine job durations in the "Timing of the Transition to Stable Employment" subsection.

	N			Mean			25th Percentile			50th Percentile			75th Percentile
Age	B	H	W	B	H	W	B	H	W	B	H	W	B	H	W
a. High school dropouts
16	328	246	558	0.0	0.0	0.1	0	0	0	0	0	0	0	0	0
17	325	244	553	0.2	0.4	0.6	0	0	0	0	0	0	0	0	1
18	324	242	540	0.8	1.4	1.5	0	0	0	0	1	1	1	2	2
19	322	239	532	1.7	2.4	2.7	1	1	1	1	2	2	2	4	4
20	319	235	528	2.4	3.3	3.8	1	2	2	2	2	3	3	4	5
21	316	233	525	3.1	4.0	4.7	2	2	3	3	3	4	4	4	6
22	312	231	516	3.9	4.9	5.6	2	2	3	3	3	5	5	5	7
23	307	224	509	4.5	5.7	6.3	2	2	4	4	4	6	6	6	9
24	300	216	501	5.2	6.4	7.1	3	3	4	5	5	6	7	8	10
25	296	210	488	5.8	7.0	7.7	4	3	5	5	6	7	8	8	10
26	266	194	442	6.5	7.6	8.4	4	4	5	6	6	8	9	9	10
27	208	133	321	7.0	8.4	9.1	4	4	5	6	7	9	9	10	10
28	146	--	179	7.3	--	9.0	4	--	5	6	--	9	10	--	10
b. High school graduates
16	373	158	696	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0
17	372	158	695	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0
18	368	157	692	0.3	0.4	0.4	0	0	0	0	0	0	0	1	1
19	363	155	684	1.1	1.4	1.6	0	1	1	1	1	1	2	2	2
20	362	153	678	2.0	2.3	2.6	1	1	1	2	2	2	3	3	3
21	358	153	668	2.7	3.2	3.4	2	1	2	2	3	3	4	4	5
22	356	151	661	3.3	3.8	4.2	2	2	2	3	3	4	4	5	6
23	354	149	657	4.0	4.4	4.8	2	2	3	4	4	4	5	6	6
24	347	146	650	4.6	4.9	5.5	3	2	3	4	5	5	6	7	7
25	339	144	642	5.2	5.4	6.0	3	3	3	5	5	5	7	7	8
26	317	134	584	5.8	5.8	6.5	3	3	3	5	5	6	8	8	9
27	246	115	456	6.2	6.4	6.8	4	3	3	6	6	6	8	9	9
28	185	--	334	6.8	--	7.3	4	--	4	7	--	6	9	--	10
29	118	--	207	7.5	--	7.7	4	--	4	7	--	7	10	--	10
c. Some college
16	195	121	416	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0
17	194	121	414	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0
18	193	121	413	0.0	0.1	0.0	0	0	0	0	0	0	0	0	0
19	191	121	412	0.3	0.4	0.3	0	0	0	0	0	0	0	1	0
20	191	119	411	0.8	1.1	1.0	0	0	0	0	1	1	1	2	2
21	187	118	402	1.5	1.9	1.9	0	1	1	1	2	2	2	3	3
22	187	117	396	2.3	2.9	2.8	1	1	1	2	3	2	3	4	4
23	186	116	391	3.1	3.6	3.6	1	2	2	3	3	3	4	5	5
24	186	114	387	3.8	4.2	4.5	2	2	2	3	3	4	5	6	6
25	181	108	383	4.4	4.9	5.1	2	3	3	4	4	5	6	7	7
26	169	102	366	5.0	5.7	5.7	3	3	3	4	5	5	6	7	8
27	137	--	298	5.1	--	6.3	3	--	4	4	--	6	7	--	8
28	111	--	246	5.6	--	6.8	3	--	4	5	--	6	8	--	9
NOTES: B = black non-Hispanic, H = Hispanic, W = white non-Hispanic and other.      A value of 10 indicates 10 or more jobs.      N is the number of individuals in the sample at least through a given age. Results are shown when sample size for a given age-SLG-race/ethnic group combination exceeds 100.

Patterns for Women

Table 4.3 reports the mean number of jobs and distribution of the number of jobs for the NLS-Y women. Consistent with other studies (e.g., BLS, 1992; Veum and Weiss, 1993), the data reveal that women in the two lowest SLGs hold fewer jobs on average than men at each age. For example, at age 28, high school dropout men have accumulated nearly 1 additional job compared with women (8.6 versus 7.7 jobs); a similar gap exists for high school graduates (7.2 versus 6.3). The gap with men is smaller for those with some college, whereas college graduate women accumulate jobs more quickly compared with men in the same SLG.[19]

Compared with SLG differences between men, the differences between women in the four SLGs is smaller when viewed in terms of years since school leaving. For example, approximately six years after school leaving (at age 29), college graduate women have held 4.7 jobs. Women with some college and those who graduated from high school, who have approximately the same potential experience at age 26 and 25, respectively, have held 5.1 and 5.2 jobs on average, respectively. The dropout group at about the same point in the work history (at age 24) has held 5.7 jobs. Men in the same four SLGs (from dropouts to college graduates) at the same ages have held 6.7, 5.9, 5.6, and 3.8 jobs, respectively, a gap of nearly 3 jobs between the dropouts and college graduates. Within each SLG, there is less heterogeneity among women than among men, as seen by the narrower spread between the number of jobs held at the 25th versus the 75th percentile.

Timing of the Transition to Stable Employment

The results on the number of jobs held suggest a considerable amount of milling about--i.e., not holding any job for very long. For a given number of years since school leaving, the amount of milling about is lower for those in the higher schooling groups. Within each SLG, the number of jobs held is almost always higher for whites than for blacks, and, sometimes, than for Hispanics. These differences could indicate either shorter job durations for young white men or, alternatively, more time spent out of work for young black men. Likewise, while women who left school at the same point typically hold fewer jobs on average when compared with men at each age, it is not clear that this difference indicates a more stable transition in the early career.

	N				HSDO				HSG				SC				CG
Age	HSDO	HSG	SC	CG	Mean	25th	50th	75th	Mean	25th	50th	75th	Mean	25th	50th	75th	Mean	25th	50th	75th
17	812	1306	1000	378	0.5	0	0	1	0.0	0	0	0	0.0	0	0	0	0.0	0	0	0
18	803	1301	999	377	1.3	0	1	2	0.5	0	0	1	0.0	0	0	0	0.0	0	0	0
19	793	1290	995	376	2.1	1	2	3	1.5	1	1	2	0.3	0	0	0	0.0	0	0	0
20	786	1277	991	374	3.0	1	3	4	2.4	1	2	3	0.9	0	1	1	0.0	0	0	0
21	782	1267	986	374	3.8	2	3	5	3.1	2	3	4	1.7	0	1	2	0.1	0	0	0
22	776	1253	978	370	4.5	2	4	6	3.6	2	3	5	2.5	1	2	4	0.7	0	0	1
23	769	1236	967	367	5.1	3	5	7	4.2	2	4	6	3.3	2	3	4	1.7	1	1	2
24	760	1222	955	364	5.7	3	5	8	4.7	3	4	6	3.9	2	3	5	2.4	1	2	3
25	747	1198	940	359	6.2	4	6	8	5.2	3	5	7	4.5	2	4	6	2.9	1	2	4
26	698	1121	903	348	6.8	4	6	9	5.6	3	5	7	5.1	3	4	7	3.4	2	3	5
27	500	877	783	309	7.3	4	6	10	6.0	3	6	8	5.5	3	5	7	3.8	2	3	5
28	302	613	660	272	7.7	5	7	10	6.3	4	6	9	5.9	3	5	8	4.3	2	4	6
29	--	366	529	236	--	--	--	--	6.5	4	7	9	6.3	4	5	8	4.7	2	4	6
30	--	171	365	184	--	--	--	--	6.2	4	7	10	6.4	4	6	8	4.9	3	4	8
31	--	--	222	--	--	--	--	--	--	--	--	--	6.4	4	6	9	--	--	--	--
NOTES: A value of 10 indicates 10 or more jobs.     N is the number of individuals in the sample at least through a given age. Results are shown when sample size for a given age-SLG combination exceeds 150.

To explore this issue further, we continue our dynamic analysis by examining the time since school leaving when young men and women first hold a job lasting one, two, or three years.

We view the experience of holding a job for one to three years as one useful measure of the concept of stable employment. It is the opposite of milling about. While we do not evaluate whether these are "good jobs" on the basis of compensation or potential for career advancement, tenure on the job is one measure of the process of settling down and a possible indicator of the transition to a career job. We also examine the sensitivity of our results to alternative definitions of job duration that have been used in the literature.

Patterns for Men

Figures 4.1, 4.2, and 4.3 plot, for each year since school leaving, the percentage of men at each age ever in a job that lasted one, two, or three years, respectively. For example, five years after school leaving, Figure 4.3 shows that about 21 percent of high school dropouts have ever held a job that lasted three years, whereas 55 percent of college graduates have done so. About one-third of those in the high school graduate and some college groups have achieved the same status.

This general pattern exists for each measure of job duration. In all cases, measuring time from school leaving, those in the college graduate group make the transition to stable employment fastest; high school dropouts make the transition the slowest. In the middle are male high school graduates and those with some college. The behavior of these latter two groups (as measured by time since leaving school) is nearly indistinguishable.

The preceding figures have plotted the transition to stable employment by time since school leaving. Figures 4.4, 4.5, and 4.6 and Table 4.4 display the same information by the age of the individual. Shortly after school leaving, the age at school leaving dominates these figures: High school dropouts are in the labor market for the longest period of time, so they have more time to experience a job lasting one, two, or three years. This effect wears off quite quickly. Within a year or two after it is chronologically possible, high school graduates overtake high school dropouts in terms of the percentage who have ever held a job for one, two, or three years. By age 30, college graduates rise to the level of the other three groups. For longer jobs, college graduates overtake those with some college and dropouts within a year or two of the first possible time (i.e., four or five years after they enter the labor market). While those with some college eventually overtake high school dropouts, through age 30 they tend to lag behind high school graduates.

Figure 4.1--Percentage of Men Ever in a Job 1 or More Years, by Years Since School Leaving

Figure 4.2--Percentage of Men Ever in a Job 2 or More Years, by Years Since School Leaving

Figure 4.3--Percentage of Men Ever in a Job 3 or More Years, by Years Since School Leaving

Figure 4.4--Percentage of Men Ever in a Job 1 or More Years, by Age

Figure 4.5--Percentage of Men Ever in a Job 2 or More Years, by Age

Figure 4.6--Percentage of Men Ever in a Job 3 or More Years, by Age

For college graduates, there is relatively little variability in the timing of the transition to stable employment. The median male college graduate holds a job for one year before his 24th birthday, a job for two years before his 26th birthday, and a job for three years before his 27th birthday (see panel d of Table 4.4). A college graduate male at the 25th percentile of the duration distribution also holds his job lasting one year before age 24, and essentially stays with that job, progressing to each later cutoff (two years and three years) about a year later. Just before age 25, a college graduate male at the 75th percentile holds a job for one year. He takes three more years to stay in a job two years, and has not stayed in a job three years by age 29, when the data become too sparse for analysis. The pattern for the median male with some college is similar to that for the median college graduate. There is more heterogeneity because of the variation in date of leaving school and the fact that some people return to school.

At the other extreme, despite the fact that he usually entered the labor force before his 17th birthday, the median male dropout does not hold a job for a year until just after his 20th birthday (see panel a of Table 4.4). The median dropout does not reach the 2-year- and 3-year-tenure points until ages 23 and 26, respectively. This implies that the median dropout did not enter a job lasting one, two, and three years until he was 19, 21, and 23, respectively.[20] For high school dropouts, the variance is substantial. At the 25th percentile of the job-duration distribution, dropouts reach the 1-, 2-, and 3-year-tenure points before ages 19, 21, and 23, a pace that is faster than that of the median high school graduates. Dropouts at the 75th percentile do not reach the 1-year-tenure point until age 22, and the 2-year-tenure point is just reached by age 29, when the data become too sparse for further analysis.

Finally, we turn to the pattern for high school graduates, the focus of the frequently cited stylized facts concerning the school-to-work transition. The median high school graduate has held a job for at least one year by the time he turns 21, two years by the time he turns 23, and three years by the time he turns 26 (see panel b of Table 4.4). Subtracting the time required to achieve each job tenure, we conclude that the median male high school graduate entered a job that would last one, two, and three years by the time he was 19, 20, and 22, respectively.[21]

If holding a job for two or even three years is not "milling about," then the patterns in Table 4.4 do not support the impression conveyed in the CSAW (1990) report that the typical high school student mills about in the labor market until age 23 or 24. It is true that the median high school graduate does not settle immediately into a long-tenure job. However, characterizing the settling-down process as lasting into the mid-twenties (e.g., 24 or 25) is overly pessimistic for the typical male in that group.

We reach this conclusion for a group of high school graduate men that includes those who returned to school full time (nearly 30 percent of the sample). If we exclude those who returned to school, the median transition to stable employment occurs somewhat earlier (see Appendix C). For example, the median high school graduate who never returned to school first enters a 3-year-tenure job at age 21 versus age 22 for his counterpart at the median of the group who ever returned to school.

While this is the pattern for the median high school graduate, the experience varies considerably at the extremes of the distribution. At the 25th percentile, high school graduates entered the 1-, 2-, and 3-year jobs while they were 18, 18, and 19 years old, respectively--that is, from one to three years ahead of the median. This probably describes what is possible in the U.S. system for young men with "successful" transitions. At the 75th percentile, young male high school graduates experienced a school-to-work transition that corresponds more closely to the perception noted at the beginning of the report. At that point in the distribution, graduates entered the 1-, 2-, and 3-year jobs at 20, 23, and 25, respectively. These transitions correspond more closely to the portrait conveyed in the CSAW (1990) report.

A Sensitivity Analysis

These results provide a considerably brighter picture of the school-to-work transition than has been presented elsewhere in the literature using the same data (see, for example, Osterman and Iannozzi, 1993). The difference derives primarily from different methods of summarizing dynamic labor market data. We define "milling about" as ending permanently when a young person first enters a job that will ultimately last more than M years. In our discussion, we subtracted M years from the age at which we observed that the individual has been in the job M years. We then considered the resulting age (the age at which he or she first entered the job that would last M years) as the end of "milling about." At least two other concepts are possible. First, we could ask whether or not the job a person is currently in has lasted or will last at least M years. Second, we can ask whether the current job has already lasted at least M years (Osterman and Iannozzi, 1993, employ this last definition).

For high school graduate men, Table 4.5 (panel b) compares the results from these differing concepts as of the time a person is exactly a given age (on each birthday) for job durations of one, two, and three years. (For purposes of comparison, Table 4.5 also presents results for the three other SLGs. The qualitative results are similar.) For each age and job duration, we present three numbers corresponding to the three concepts just described. The column labeled "L" corresponds to the longest job a person has ever held as of the given age (our preferred definition). The column labeled "E" corresponds to the eventual length of the current job at that age. The column labeled "C" corresponds to the length to date of the currentjob at that age. There is a formal relation between these concepts. The current job duration is always less than or equal to the eventual duration of the job, which is always less than or equal to the duration of the longest job M years later (i.e., M rows down the table in the "L" column).

According to Table 4.5, the tenure on the current job clearly gives the most negative results (column C). Consider, for example, the 2-year-duration job as of age 26 for high school graduate men (panel b of Table 4.5). More than half of all high school graduates at age 26, 54.6 percent (100.0 - 45.4), have not been in their current job for even two years. Note, however, that for nearly half, 46.5 percent [(70.8 - 45.4)/54.6], of those people whose current job has not lasted two years, the current job itself will last two years. Further, nearly one-quarter, 22.6 percent [(77.4 - 70.8)/(100.0 - 70.8)], of the men whose current job will not last two years have already held a job that has lasted two years. Put differently, half of all high school graduates at a given age are not in a job that will last two years until nearly age 27. However, just after his 21st birthday, the median high school graduate is in a job that will last at least two years. Finally, before his 20th birthday, the median high school graduate has, at some earlier point in his work history (perhaps not the current job), entered a job that eventually lasted at least two years.

Following our earlier discussion of the literature on job matching (e.g., Mincer and Jovanovic, 1981; Flinn, 1986; McCall, 1990), we are reluctant to view all job turnover as bad.

The literature on job matching suggests that most job changes involve sizable wage increases. From this perspective, we are concerned about measures of employment stability that consider to be a poor outcome (i.e., an indication of an unsuccessful transition to stable employment)[22] those individuals who have not, at an arbitrary point in time, been in their current job for several years. A similar criticism applies to definitions that consider the eventual duration of the current job.

Nevertheless, we agree that in and of itself, failure to stay on a job for a significant period of time (one, two, or even three years) often indicates some problem. If employers do not expect young workers to stay on the job even for such a moderate period of time, they will not invest in even the minimal training required for primary-sector jobs. However, a worker who spends several years at one employer and then moves on to a new job (often with a large increase in pay) is not a failure: The new job could also last several years.

Even a short job (under a year) between two longer jobs need not be a failure. Perhaps the short job did not "work out"; perhaps it was deliberately temporary until an appropriate career-enhancing job became available (perhaps the worker had already lined up that next good job). For all these reasons, we prefer our definition of transition period as the time until the young worker first enters a job that will eventually last more than M years. And, again for the same reasons, we are concerned that the alternative definitions we have discussed present an overly pessimistic view of labor market dynamics.

Differences by Race/Ethnicity

The analysis for all young men indicated that there is considerable variation across SLGs in the timing of the transition to stable employment. Further, there is heterogeneity within these groups as evidenced by differences in the experiences of the typical male versus those who make the transition more rapidly or more slowly. We would like to know if there is variation, as well, among men in different race/ethnic groups. To address this question, in Table 4.6 we report the percentage of black, Hispanic, and white men within each SLG who have held a job of one, two, and three years' duration at each age. For the 2-year-duration job, Figures 4.7, 4.8, and 4.9 plot the percentage of men in each of three groups--high school dropouts, high school graduates, and those with some college, respectively[23]--attaining this status.

Figure 4.7--Percentage of High School Dropout Men Ever in a Job 2 or More Years, by Age

Figure 4.8--Percentage of High School Graduate Men Ever in a Job 2 or More Years, by Age

Figure 4.9--Percentage of Men with Some College Ever in a Job 2 or More Years, by Age

While there is considerable heterogeneity between and within SLGs, there is surprisingly little variation in the distribution of job durations for men of different race/ethnicity within each SLG. Focusing on Figures 4.8 and 4.9, we find it hard to distinguish between the experience of white, black, and Hispanic high school graduates or those with some college. At each age with sufficient data to make a comparison, the fraction who had held a job for at least two years is similar for all three groups. The median black and white males with some college, for example, each had held a job for at least two years before their 26th birthday; the median Hispanic male had reached this status in the months before turning 25.

Among high school graduates, more diversity is indicated in the experience among black males than that among their white and Hispanic counterparts. Although the lines in Figure 4.8 track each other closely until age 23, they diverge somewhat after that point. This divergence can be viewed in terms of the differential experience of men at the 50th versus the 75th percentile of the distribution. For each of the three groups, the median high school graduate has held a job with at least two years of tenure before turning 23. In contrast, the white male high school graduate at the 75th percentile reaches a job with 2-year tenure before turning 26, and the Hispanic male at the same point in his distribution reaches this status several months earlier. The black male high school graduate at the 75th percentile has not held a job for two years until just before his 27th birthday. Thus, while the typical (i.e., at the 50th percentile) white, black, and Hispanic high school graduates show a similar transition to stable employment, at the tails of the distribution, black males make the transition more slowly.

Although the patterns are largely similar among high school graduates and men with some college, differences among black, Hispanic, and white male high school dropouts are considerably more pronounced. They are evident in Figure 4.7, which shows that black male dropouts make the transition to stable employment at a slower pace than Hispanic or white male dropouts. While there is evidence that Hispanic men lag behind their white counterparts, the differences between these two groups are not as sharp as the contrast with the pattern for black dropouts. For example, the median white and Hispanic males have held a job for a least two years by the time they reach 24, whereas this status is not attained by black dropouts until more than one year later. A similar pattern is evident for the 1-year-tenure job, and will also hold for the 3-year-tenure job (for which the pattern must be inferred after age 28).

The patterns among race/ethnic groups in the transition to stable employment are surprising in light of the preceding results regarding the number of jobs held early in the career. Recall, for example, that white high school dropouts at each age held more jobs on average than their Hispanic counterparts, who in turn held more jobs than black high school dropouts. This result, however, does not translate into a more difficult transition to stable employment for white men. In fact, the reverse is true: A higher fraction of white high school dropouts has held a job for at least two years at each age compared with black dropouts. Likewise, the higher number of jobs held on average by white or Hispanic high school graduates (and those with some college) compared with blacks does not result in large differences between the race/ethnic groups in the timing of the transition to stable employment.

It is possible to reconcile the results of the analysis of the number of jobs held with our findings on job durations. If black men are more likely to experience spells between jobs, during which they are in school, unemployed, or out of the labor force, they will hold fewer jobs as they make the transition to stable employment. Conversely, if white and Hispanic men are more likely to make transitions from job to job, with many jobs lasting for only a short duration, they will hold more jobs in their early career. The static analysis in Section 3 supports this hypothesis: Black men, especially dropouts, are much more likely at any point in time to be unemployed or not in the labor force. Thus, it is unlikely that variation in the propensity to return to school can explain the racial differences we observed. We leave reconciling these dynamic differentials to subsequent work.

Patterns for Women

Consistent with our earlier analyses, the patterns for women vary in a number of ways from those evident for men. Table 4.7 shows the percentage of women at each age and SLG who have ever held a job for one, two, or three years; Figures 4.10, 4.11, and 4.12 plot the percentages for the four SLGs for each successive job duration. Like the men, young women in higher SLGs make the transition to stable employment more rapidly. Most notably, women leaving school with a college degree or with some college overtake the less educated groups much more rapidly compared with the men. This difference is due to the speed with which high school dropout and graduate women make the transition compared with their male counterparts.

In fact, regardless of the measure of job duration, women high school graduates make the transition even more slowly than high school dropout men. However, college graduate women and those with some college have patterns that closely mirror those of men in the first few years after leaving school--perhaps because they delay childbearing.

Figure 4.10--Percentage of Women Ever in a Job 1 or More Years, by Age

Figure 4.11--Percentage of Women Ever in a Job 2 or More Years, by Age

Figure 4.12--Percentage of Women Ever in a Job 3 or More Years, by Age

For college graduate women, then, the dynamic process in the early labor market career, whether measured by the number of jobs held or the timing to stable employment, closely matches the experiences of men in the same SLG. At the other end of the spectrum of school leaving, women dropouts and high school graduates face a slower transition to stable employment compared with men, a result that is not reflected in a higher number of jobs held, but in more time spent out of the labor force or looking for work.

The Dynamic Perspective of the School-to-Work Transition

The dynamic analysis presented in this section allows us to examine more closely the complexities of the early labor market experience of young men and women in the NLS-Y. The early career is inherently a dynamic period, with transitions in and out of the labor force, and between jobs of various durations. The employment histories provided in the NLS-Y reveal that this period can be characterized as one in which numerous jobs are held. While young men do not immediately enter jobs that will last for a considerable length of time, such as two or three years, that transition is not delayed until the mid-twenties. The typical, or median, high school graduate who at age 26 (about 8 years after leaving school) has held 6 jobs, has held at least one of those jobs for at least three years. Thus, while he is age 21, three years after leaving high school, the median high school graduate who does not return to school has already entered a job that will last three years.

Of course, the pattern observed for the typical male obscures the extent of the differences that exist between youth who make this transition more rapidly or more slowly. The analysis in this section reveals that the time until a job of one, two, or three years' duration is attained occurs much sooner after leaving school for those who attain more schooling. High school dropouts clearly fare worse in the process, and minority men in this group do even worse. Surprisingly, differences across race/ethnic groups within each SLG for men were less pronounced than the large differences that persist among the SLGs. While women college graduates experience a transition to stable employment that is similar to that of their male counterparts, less educated women make the transition even more slowly than men through more time spent in non-employment.

[15]Our results will be different from those of other recent studies, which measure all jobs held starting at a certain age whether or not the youth had left school, and classify youth by schooling attainment as of the most recent interview date (rather than at school leaving).

[16]Our estimates of the mean number of jobs held for higher SLGs are lower than the estimates provided by other analyses. The reason appears to be that other authors (e.g., BLS, 1992, 1993; Veum and Weiss, 1993) count all jobs since age 18, whereas we count only jobs since school leaving. For this reason, we show fewer jobs at any given age for higher SLGs, while other authors (e.g., BLS, 1992, 1993; Veum and Weiss, 1993) find the reverse relationship.

[17]A similar pattern is reported by Veum and Weiss (1993) in their analysis of the average number of jobs held starting from age 18 by men and women in groups defined by completed schooling and race/ethnicity.

[18]Veum and Weiss (1993) report that black male dropouts have fewer weeks of work experience and more weeks of unemployment at each age between 18 and 27 compared with their white or Hispanic counterparts. A similar gap is evident for high school graduates, while the differences are almost eliminated for higher education groups. Their analysis, unlike ours, classifies youth by their completed schooling rather than SLG.

[19] When only full-time jobs are counted, college graduate women hold fewer jobs on average at each age than men in the same SLG.

[20]This computation proceeds by noting that if more than 50 percent of all people in an SLG have been in a job Myears by the birthday when they turned Ayears, then the median person reached that point when he was A- 1 years old, and he entered the job at least A- 1- M years earlier. So, for example, reading from Table 4.4, the high school dropout percentiles for 2 years' job tenure are 46.6 and 54.1 at the birthdays at which the young men turned 23 and 24, respectively. Thus, the median male reached his 2-year-tenure date while he was 23 and had entered a job that would eventually last at least two years by the time he was 21.

[21]See footnote 6 for the computation.

[22]This distinction--time in a job at an interview versus completed job duration--is discussed for the case of job tenure in Horvath (1982), which is based on a discussion in the unemployment literature (Bowers, 1980).

[23]The sample sizes become too small for a comparison by race/ethnicity within the college graduate group.