Cumulative Exit Rates From Job And Employment Spells
For Those Starting Low-Wage Jobs, By Gender
Note: All figures were calculated using the longitudinal panel weight.
[ Main page of Report | Contents of Report ]
Thus far, we have examined the overall employment experiences and wage growth of low-wage earners over a three-and-one-half-year period after job start. For these analyses, the worker was the unit of analysis, and we examined aggregate measures of potentially discontinuous employment and nonemployment spells that workers experienced over the fixed follow-up period. Another interrelated way to examine the labor market experiences of low-wage workers is to directly examine the duration of their employment and nonemployment spells. For these analyses, the spell, rather than the worker, is the unit of analysis.
These spell analyses allow us to address the following important study questions:
We addressed these questions using information on the duration of job, employment, and nonemployment spells that started during the panel period. We used life table statistical methods to examine spell durations for the full sample, by gender, and for key subgroups of low-wage workers.
Our spell analysis paints a complex picture of the labor market dynamics of low-wage workers. Most importantly, we find that the job, employment, and nonemployment spells of low-wage workers during the mid- to late 1990s were short, and that there was substantial diversity in the ways in which these spells ended. For instance, the median duration of low-wage job spells was about four months for both males and females; about 80 percent ended within a year, and more than 90 percent ended within two years. About 39 percent of male low-wage workers and 28 percent of female low-wage workers exited their low-wage jobs directly into higher-wage employment within three-and-one-half years after job start; at the same time, however, 31 percent of spells for males and 41 percent of spells for females ended in nonemployment (with the remainder of spells ending in another low-wage job). Similarly, more than one-half of those who exited their low-wage jobs into higher-wage jobs returned to the low-wage labor market within two years, and about 87 percent of males who exited their low-wage jobs into nonemployment became reemployed within two years (with one-quarter entering high-wage jobs and the remainder entering low-wage jobs).
These results suggest that job mobility was very common; many workers bounced in and out of the low-wage and higher-wage labor markets.(45) Furthermore, our results indicate that the pathways that led to general improvements in economic prospects over time (discussed in the overall employment and wage progression analyses) differed significantly across workers and were not smooth for most workers. Finally, and not surprisingly, we find that the same subgroups of workers who typically had the best overall employment experiences and wage growth also had the best spell-related outcomes.
[ Go to Contents ]
We conducted multiple spell analyses to examine exit rates out of low-wage jobs and reentry rates into the low-wage labor market. For each spell analysis, the sample contains an entry cohort of job, employment, or nonemployment spells that started during the panel period. Thus, an individual could contribute more than one spell to an analysis file.
Each analysis file contains one observation per month of the spell. We constructed a dependent variable that was set to zero in months when the spell was in progress, and to 1 when the spell ended (or in some analyses, to positive codes signifying the type of exit or reentry). The last observation for a spell corresponds to the month when the spell ended, or to the end of the panel period for spells that were still in progress at that time (that is, for right-censored spells). The analysis files also contain individual and job characteristics pertaining to the month in which the spell started that were used for the subgroup analysis.
Next, we discuss the various types of spells that we examined and the life table procedures that we used to estimate spell durations.
A central, and complicated, analytic issue is how to define job, employment, and nonemployment spells (that is, the rules used to assign zeros and positive codes to the dependent variables discussed above). To facilitate this discussion, we first list the five possible states into which a low-wage worker could exit:
Using these possible exit states, we conducted duration analyses for four types of job and employment spells, each of which addresses a slightly different analytic question:
Similar procedures were used to construct spells for those who began medium- and high-wage jobs during the panel period.
We examined two types of spells for our analyses of reentry into the low-wage labor market. First, we examined the rate at which those who exited their low-wage jobs into nonemployment (that is, into exit states 4 and 5) returned to the low-wage and higher-wage labor markets. Second, we examined the extent to which those who exited their low-wage jobs into higher-wage jobs returned to the low-wage sector.
To examine the duration of job, employment, and nonemployment spells, we used "life table analyses." Spells can be broken down into months; for each month, the life table displays the estimated hazard rate and cumulative exit rate. The hazard rate is the probability that a spell ended in a particular month, given that the spell lasted at least until the beginning of that month. The cumulative exit rate, obtained from the estimated hazard rates, is the unconditional probability that a spell ended within a given number of months. The cumulative exit rate enables policymakers to answer such questions as: Of the next 100 people who begin a low-wage job spell, how many will exit their low-wage jobs within one year?
A major advantage of using life table methods is that they can effectively treat right-censored spells (that is, spells still in progress at the end of the observation period). Right-censored spells contribute information to the life table up to the month in which they are right-censored (that is, up to the time we no longer have information on them). For example, if a spell is right-censored 12 months after the spell started, then that spell is included in the hazard rate calculations (that is, enters the denominator of the calculations) for months 1 to 12, but not afterward.
The treatment of left-censored spells (that is, spells in progress at the start of the panel) is more problematic, because the duration distributions of left-censored and non-left-censored spells are likely to differ. For example, suppose a low-wage job spell started one year prior to the start of the panel period. Then, that spell would be observed in the data only if it lasted longer than one year (it would not be observed if it ended prior to the panel period). Furthermore, counting from month 1 of the panel period, the spell is likely to last longer than a typical non-left-censored low-wage spell because of duration dependence (that is, spell exit rates often decrease the longer the spell has been in progress). Thus, left-censored spells are likely to be longer on average and to have a different duration distribution than are typical spells.
Left-censored spells, however, can be included in the life table analysis, because the wave 1 core files contain information on the start dates of left-censored spells. The left-censored spells contribute information to the life table starting in the month in which they are left-censored. For example, a spell that had been in progress for 12 months would enter the life table starting in month 12. This procedure, however, produces unbiased estimates only if we assume a stationary environment (that is, if spell duration distributions did not change over time). This assumption, however, may be unrealistic for spells that had been in progress for a long time due to changes in labor market structure and conditions. Furthermore, because SIPP does not contain prepanel information on hourly wages, left-censored spells can be included in the analysis only if we assume that left-censored low-wage jobs were low-wage jobs for the entire period between job start and month 1 of the panel period.
For these reasons, we excluded left-censored spells in our main spell duration analyses (the approach that most researchers conducting event history studies use). However, left-censored spells were included in some analyses to examine issues pertaining to the duration of longer spells than could be observed in the panel period and to check the robustness of study findings.
The life table methods described above can be extended to examine the rate at which workers leave the low-wage labor market, by type of exit. In this "competing risks" framework, the dependent variable for the analysis was set to zero in months the spell was in progress and to a positive code--signifying the specific exit type--in the month the spell ended. Thus, spells contributed information to the life table up to the month that they ended (that is, until a positive code appeared) or until the end of the panel period for right-censored spells. In this framework, the estimated monthly hazard and cumulative exit rates across the exit types sum to the corresponding values for the overall spell analysis where we did not distinguish between exit types.
The life tables themselves contain a great deal of information and can be complicated. Because the cumulative exit rates efficiently and intuitively summarize the life table results, our presentation focuses on them. Furthermore, when presenting results for the subgroup analyses, we present summary information such as the median spell duration, as well as the percentage of spells that ended within a given number of months. We also conducted statistical tests to gauge whether the spell duration distributions differed across levels of a subgroup using the log-rank statistic.(46) All statistics were constructed using the longitudinal panel weight.
Finally, for several reasons, we present life table results by wave only (that is, in four-month intervals from 4 to 44 months after job start). First, as discussed in the Methodological Appendix, the constructed hourly wage for a particular job or business was constant within a wave. Second, sample members tended to report being employed (or unemployed) for the entire wave rather than for only specific months covered by the wave. Consequently, we find more changes in low-wage job status across waves than within waves, so that the estimated hazard rates spike at the "seam" points. Thus, we present the life table results in four-month intervals only.
The sample contains a large number of low-wage job spells (Table VI.1). The larger number of spells for females than males (10,259 spells for 5,985 female workers, compared to 6,373 spells for 3,934 male workers) is consistent with our earlier findings that low-wage workers are disproportionately female. About 20 percent of spells are right-censored, and nearly 30 percent are left-censored. Few are both right- and left-censored. Because of duration dependence, mean observed spell durations are considerably longer for left-censored spells than for non-left-censored ones.(47)
| Spell Type for Males | Spell Type for Females | |||
|---|---|---|---|---|
| Job | Employment | Job | Employment | |
| Low-Wage Spells | ||||
| Total Number of Spells | 6,373 | 4,882 | 10,259 | 7,755 |
| Number of Spells per Worker (Percentages) | ||||
| 1 | 62 | 75 | 58 | 73 |
| 2 | 22 | 18 | 23 | 20 |
| 3 or more | 16 | 7 | 19 | 7 |
| (Average number) | (1.7) | (1.3) | (1.8) | (1.4) |
| Percentage of Spells That Are: | ||||
| Right-censored | 18 | 22 | 20 | 25 |
| Left-censored | 29 | 38 | 28 | 36 |
| Right- and left-censored | 4 | 6 | 3 | 6 |
| Mean Observed Spell Duration (Months)(a) | ||||
| Non-left-censored spells | 7 | 8 | 8 | 10 |
| All spells | 25 | 31 | 25 | 32 |
| Percentage of Low-Wage Spells with Exit Type(b) | ||||
| Another low-wage job | 18 | NA | 21 | NA |
| Medium- or high-wage job | 32 | 22 | ||
| In the same job | 21 | 15 | ||
| In a different job | 11 | 7 | ||
| Unemployment | 14 | 10 | ||
| Not in the labor force | 13 | 21 | ||
| Spells of Any Wage Type | ||||
| Total Number of Spells | 6,170 | 3,943 | 10,057 | 6,832 |
| Number of Spells per Worker (Percentages) | ||||
| 1 | 61 | 77 | 58 | 74 |
| 2 | 22 | 16 | 23 | 19 |
| 3 or more | 17 | 7 | 19 | 7 |
| (Average number) | (1.7) | (1.3) | (1.8) | (1.3) |
| Percentage of Spells That Are: | ||||
| Right-censored | 32 | 53 | 32 | 48 |
| Left-censored | 28 | 42 | 27 | 39 |
| Right- and left-censored | 10 | 27 | 9 | 21 |
| Mean Observed Spell Duration (Months)(a) | ||||
| Non-left-censored spells | 10 | 13 | 10 | 13 |
| All spells | 29 | 60 | 28 | 41 |
| Source: 1996 SIPP longitudinal
files for those in low-wage jobs. Note: All figures are unweighted. A job spell of any wage type pertains to the period that the worker was employed with the initial employer, while the employment spell of any wage type includes continuous changes from one job to another. Low-wage job spell pertains to the duration with the initial employer, in which the worker continues to receive low pay, and the low-wage employment spell includes continuous changes from one low-wage job spell to another. The definitions for each spell type are given in Section A.1. a. Figures pertain to the mean spell length observed during the panel period, including spells that are still in progress at the end of the period (that is, right censored spells). Thus, the figures are shorter than the ultimate mean lengths of the spells. b. Figures pertain to exit types for non-left-censored spells only. NA: = Not applicable |
||||
The analysis file contains multiple low-wage job spells for a substantial number of workers (Table VI.1). On average, male and female workers each contributed about 1.8 spells to the file, and about 40 percent contributed at least 2 spells. These results are consistent with our findings from the overall employment analysis that many low-wage workers exit low-wage jobs, but many return to the low-wage labor market.
Exit types vary across low-wage workers (Table VI.1). The most common exit type for both male and female low-wage workers in our sample was into higher-paying jobs. Among non-right-censored spells, about 32 percent of spells for males and 22 percent of spells for females ended in this way. Furthermore, most of these spells ended in a higher-wage job with the same employer rather than with a different employer. At the same time, however, many workers exited their low-wage jobs into another low wage job (20 percent of spells for males and females) or into nonemployment.
Interestingly, spell information for low-wage job and low-wage employment spells are similar (Table VI.1). This occurs because only a relatively small percentage of workers moved directly from one low-wage job to another.
The sample contains fewer job and employment spells than low-wage job and employment spells (Table VI.1). This occurs because many low-wage job spells resulted in continued employment in higher-wage jobs. Stated differently, only a relatively small fraction of low-wage spells ended in nonemployment. Thus, mean observed spell durations are somewhat longer for the overall job and employment spells than for the low-wage spells. Similarly, a much higher percentage of overall job and employment spells are right-censored.
Finally, the analysis files contain more medium-wage than low-wage job and employment spells for both males and females (Tables E.1 and E.2). We expected these findings because our cross-sectional analysis found that the medium-wage sector is the largest labor market sector, and because our overall employment analysis found that many low-wage workers obtain medium-wage jobs. Not surprisingly, observed mean spell durations are shorter for low-wage than higher-wage job and employment spells.
[ Go to Contents ]
This section presents key findings from our life table analysis for various types of job, employment, and nonemployment spells. We present findings, by gender, for the full sample of spells, as well as for key subgroups of low-wage workers defined by their individual, household, and initial low-wage job characteristics.
a. Low-Wage Job Spells
Low-wage job spells that started during the mid- to late 1990s were typically short for both men and women (Table VI.2). About one-half of spells ended within four months after job start, about three-quarters ended within one year, and nearly 90 percent ended within two years. By 44 months after job start (the longest period for which life table results could be obtained), about 95 percent of low-wage job spells had ended. Thus, there is substantial wage and job mobility among low-wage workers.
| Month | Total | Type of Exit | ||||
|---|---|---|---|---|---|---|
| Another Low-Wage Job | Higher-Wage with the Same Employer | Higher-Wage with a Different Employer | Unemployment | Out of the Labor Force | ||
| Males | ||||||
| Number of Months After Start of Low-Wage Job | ||||||
| 4 | 51 | 12 | 13 | 8 | 9 | 9 |
| 8 | 73 | 17 | 18 | 11 | 14 | 12 |
| 12 | 81 | 20 | 21 | 12 | 15 | 14 |
| 16 | 87 | 21 | 23 | 12 | 16 | 15 |
| 20 | 90 | 22 | 23 | 13 | 16 | 16 |
| 24 | 92 | 22 | 24 | 13 | 17 | 16 |
| 28 | 94 | 23 | 25 | 13 | 17 | 16 |
| 32 | 95 | 23 | 25 | 13 | 17 | 17 |
| 36 | 96 | 23 | 26 | 13 | 17 | 17 |
| 40 | 97 | 23 | 26 | 13 | 17 | 17 |
| 44 | 97 | 23 | 26 | 13 | 18 | 17 |
| Females | ||||||
| Number of Months After Start of Low-Wage Job | ||||||
| 4 | 46 | 13 | 8 | 5 | 7 | 13 |
| 8 | 65 | 19 | 12 | 7 | 10 | 18 |
| 12 | 76 | 22 | 14 | 8 | 11 | 22 |
| 16 | 83 | 24 | 16 | 8 | 12 | 23 |
| 20 | 87 | 25 | 17 | 8 | 12 | 24 |
| 24 | 90 | 26 | 18 | 9 | 13 | 25 |
| 28 | 92 | 26 | 18 | 9 | 13 | 26 |
| 32 | 93 | 27 | 19 | 9 | 13 | 26 |
| 36 | 94 | 27 | 19 | 9 | 13 | 27 |
| 40 | 95 | 27 | 19 | 9 | 13 | 27 |
| 44 | 96 | 27 | 19 | 9 | 14 | 27 |
| Source: 1996 SIPP longitudinal
files using the entry cohort sample of 4,489 low-wage job spells for males
and 7,401 low-wage job spells for females. Left-censored spells are excluded
from the sample. Note: All figures are weighted using the longitudinal panel weight. |
||||||
Into which labor market state did low-wage workers most often exit? The answer is that there is considerable diversity in exit states, although low-wage workers most often exited into higher-wage jobs (Table VI.2). Interestingly, most of those who entered higher-paying jobs stayed with the same employer. Looking at all exits that occurred within 12 months after job start, low-wage jobs evolved into higher-paying jobs with the same employer for 21 percent of males and 14 percent of females. Over the same one-year period, an additional 12 percent of males and 8 percent of females obtained a different higher-paying job. Thus, altogether, 33 percent of male low-wage workers and 22 percent of female low-wage workers found higher-paying employment within one year. Thereafter, the cumulative exit rates into higher-wage employment leveled off to about 39 percent for males and 28 percent for females. These findings provide further evidence of some wage mobility for the low-wage population during the strong economy of the mid- to late 1990s.
At the same time, however, many workers during the mid- to late 1990s exited their low-wage jobs directly into another low-wage job or into nonemployment (Table VI.2). For instance, 27 percent of spells for females and 23 percent of spells for males eventually ended in another low-wage job. Similarly, more than one-quarter of female workers and 17 percent of male workers exited their jobs by leaving the labor force. Finally, spells ultimately ended in unemployment for about 18 percent of males and 14 percent of females. Thus, altogether, about 41 percent of spells for females and 31 percent of spells for males ended in nonemployment.
b. Low-Wage Employment Spells
Thus far, we have examined the length of low-wage job spells from the start of these spells until the worker exited into another low-wage job, a higher-paying job, or nonemployment. As discussed, we also examined the duration of low-wage employment spells, which were allowed to continue if a worker moved continuously from one low-wage job to another. Thus, these spells could end only if the worker found a higher-paying job or became nonemployed.
Low-wage employment spells tend to be slightly longer than low-wage job spells (Tables VI.2 and VI.3, the top two lines in Figure VI.1, and Tables E.3 and E.4). For example, among male low-wage workers, about 74 percent of low-wage job spells ended within one year after job start, compared to 81 percent of low-wage employment spells. The differences between the duration distributions of low-wage job and low-wage employment spells reflect the fact that about one-quarter of low-wage workers in our sample moved from a low-wage job directly into another low-wage job.
Examining the types of exits from low-wage employment spells and low-wage job spells tells a somewhat similar story (Table VI.3). As expected, transition rates into higher-wage jobs and into nonemployment are somewhat larger for low-wage employment spells (because transitioning into another low-wage job is no longer a possible exit state). For instance, about 43 percent of males eventually exited their low-wage employment spells into medium-wage jobs and an additional 6 percent exited into high-wage jobs. Thus, nearly one-half of males exited their low-wage employment spells directly into higher-paying jobs, which is somewhat larger than the corresponding figure of 39 percent for male low-wage job spells. Similarly, about 38 percent of females eventually exited their low-wage employment spells because they left the labor force, whereas the corresponding figure is 27 percent for female low-wage job spells.
These findings support our results from the overall employment and wage progression analyses that there is substantial diversity in labor market success across low-wage workers. They also support our previous findings that female low-wage workers typically have poorer labor market outcomes than male low-wage workers.
|
|
| Source: 1996 SIPP longitudinal files using
the entry cohort sample. Note: All figures were calculated using the longitudinal panel weight. |
| Total | Type of Exit | ||||
|---|---|---|---|---|---|
| Medium-Wage Job | High-Wage Job | Unemployment | Out of the Labor Force | ||
| Males | |||||
| Number of Months After Start of Low-Wage Job | |||||
| 4 | 44 | 20 | 3 | 11 | 10 |
| 8 | 65 | 28 | 4 | 17 | 15 |
| 12 | 74 | 32 | 5 | 19 | 18 |
| 16 | 82 | 36 | 5 | 21 | 20 |
| 20 | 86 | 38 | 5 | 22 | 21 |
| 24 | 88 | 39 | 6 | 23 | 21 |
| 28 | 90 | 40 | 6 | 23 | 21 |
| 32 | 92 | 41 | 6 | 24 | 22 |
| 36 | 94 | 42 | 6 | 24 | 22 |
| 40 | 95 | 43 | 6 | 24 | 22 |
| 44 | 96 | 43 | 6 | 24 | 23 |
| Females | |||||
| Number of Months After Start of Low-Wage Job | |||||
| 4 | 39 | 13 | 2 | 8 | 16 |
| 8 | 57 | 19 | 3 | 12 | 23 |
| 12 | 68 | 22 | 3 | 14 | 28 |
| 16 | 75 | 25 | 3 | 16 | 32 |
| 20 | 80 | 27 | 3 | 16 | 33 |
| 24 | 84 | 29 | 3 | 18 | 35 |
| 28 | 87 | 30 | 4 | 18 | 36 |
| 32 | 89 | 31 | 4 | 18 | 36 |
| 36 | 90 | 31 | 4 | 18 | 37 |
| 40 | 91 | 31 | 4 | 19 | 37 |
| 44 | 93 | 32 | 4 | 19 | 38 |
| Source: 1996 SIPP longitudinal
files using the entry cohort sample of 3,021 low-wage employment spells for
males and 4,926 low-wage employment spells for females. Left-censored spells
are excluded from the sample. Note: All figures are weighted using the longitudinal panel weight. |
|||||
We examined also the duration of low-wage jobs using two alternative definitions of spell end dates. First, we examined job spells, where a spell continued as long as the worker remained with their initial employer regardless of the wage received. Second, we examined employment spells, where a spell continued as long as the worker was employed in any job (regardless of the wage level received).
Not surprisingly, job spells tend to be longer than low-wage job spells (Figure VI.1 and Tables E.3 and E.4). For example, nearly 80 percent of low-wage job spells in our sample ended within one year after job start, compared to only 67 percent of job spells. Similarly, more than 90 percent of low-wage spells ended within 24 months, compared to only about 80 percent of job spells. These findings are due to the significant numbers of low-wage workers who obtained higher-wage jobs with the same employer.
Despite these findings, however, job spells are not long. About two-thirds ended within a year after job start, and more than three-quarters ended within two years. Thus, low-wage workers sometimes obtain higher-wage jobs with the same employer, but many do not remain in these higher-wage jobs for a substantial period of time.
The finding that job spells are not long, however, is not necessarily a negative result, because as discussed in the previous two chapters, among workers who were continuously employed during the follow-up period, those who switched jobs tended to have more positive labor market outcomes than those who remained with their initial employers. Thus, job turnover is an avenue for wage growth for some low-wage workers.
We also find that spell durations tend to be longer for overall employment spells than for overall job spells (Figure VI.1 and Tables E.3 and E.4). For example, one-year cumulative exit rates were about 67 percent for job spells, compared to about 51 percent for employment spells. The two duration distributions differ because a sizeable fraction of low-wage workers moved continuously from a low-wage job to another one or to a higher-paying job with a different employer.
The life table results that include left-censored spells (about 37 percent of all employment spells and 30 percent of all job spells) are similar to those that exclude these spells (Tables VI.4, E.3 and E.4). Cumulative exit rates from employment and job spells in months 4 to 44 are very similar whether or not the left-censored spells are included in the analysis (although the left-censored spells are slightly longer than their comparable non-left-censored spells).(48) These results suggest that the assumptions, discussed above, that are needed to justify the use of the left-censored spells appear to be appropriate (at least for spells that started soon before the panel period).
How long are the spells of workers who start low-wage jobs compared to those of workers who start medium- and high-wage jobs? The answer to this question can help place in perspective the life table findings for low-wage workers presented above.
To address this question, we compared two types of employment spells for those starting low-, medium-, and high-wage jobs. First, we examined the length of time workers were employed in jobs of the same wage type as their initial job. For example, we examined how long medium-wage workers remained in medium-wage jobs (either with the same employer or with a different one). Second, we examined how long workers were employed at all. For example, we examined how long those starting high-wage jobs were continuously employed in any job and at any wage level.
| Month | Males | Females | ||
|---|---|---|---|---|
| Without Left- Censored Spells |
With Left- Censored Spells |
Without Left- Censored Spells |
With Left- Censored Spells |
|
| Number of Months After Start of Low-Wage Job | ||||
| 4 | 26 | 24 | 28 | 25 |
| 8 | 42 | 39 | 42 | 39 |
| 12 | 51 | 46 | 52 | 48 |
| 16 | 57 | 52 | 58 | 54 |
| 20 | 61 | 56 | 62 | 58 |
| 24 | 64 | 59 | 66 | 62 |
| 28 | 66 | 61 | 68 | 65 |
| 32 | 69 | 63 | 71 | 67 |
| 36 | 71 | 65 | 72 | 69 |
| 40 | 72 | 67 | 75 | 72 |
| 44 | 74 | 68 | 78 | 73 |
| 48 | 70 | 74 | ||
| 52 to 104 (1 to 2 Years) | 81 | 85 | ||
| 105 to 156 (2 to 3 Years) | 85 | 91 | ||
| 157 to 208 (3 to 4 Years) | 89 | 94 | ||
| 208 to 260 (4 to 5 Years) | 94 | 97 | ||
| Source: 1996 SIPP longitudinal
files using the sample of 3,943 spells for males and 6,832 spells for females.
Note: All figures are weighted using the longitudinal panel weight. |
||||
We find that, during the mid- to late 1990s, low-wage employment spells were typically shorter than medium- and high-wage employment spells, especially for males (Figure VI.2, and Tables E.3 and E.4). Furthermore, the differences increased somewhat over time. For example, among male workers, about 74 percent of low-wage employment spells ended within one year after job start, compared to 60 percent of high-wage spells. By 24 months, differences in the cumulative exit rates were larger (88 percent for low-wage employment spells, compared to 70 percent for high-wage employment spells). This suggests that, after an initial adjustment period, higher-wage workers became more and more likely than low-wage workers to remain on their jobs. Differences in spell lengths by wage type, however, are smaller for women. The 24-month cumulative exit rate was 84 percent for females with low-wage employment spells and 76 percent for females with high-wage employment spells.
Although medium- and high-wage employment spells were somewhat longer than low-wage employment spells, they were shorter than expected. The reason is that a nontrivial percentage of medium-wage workers exited into low-wage or high-wage jobs, and a nontrivial percentage of high-wage workers exited into medium-wage jobs. Among medium-wage workers, about 30 percent of males and females ultimately exited into higher-wage jobs, and 28 percent of males and 19 percent of females ultimately exited into low-wage jobs (not shown). Among those starting high-wage job spells, about 37 percent of males and females exited into medium-wage jobs within 44 months.
|
| Source: 1996 SIPP longitudinal files using
the entry cohort sample. |
Thus, there is considerable wage movement over time for those starting medium- and high-wage spells, as well as for those starting low-wage spells. These wage fluctuations could be due partly to reporting errors in hourly wages or in monthly earnings and hours worked per week (for those who could not directly report hourly wages) or to temporary changes in labor supply effort or earnings levels (although we "smoothed" the constructed wage measures within waves to help alleviate this problem). We believe, however, that our findings reflect real movements of medium- and high-wage workers across wage categories. This interpretation is supported by findings from the overall employment analysis that many workers in our sample who started higher-wage jobs at the start of the panel period experienced multiple job and employment spells over the three-and-one-half-year follow-up period. Thus, wage and job mobility is common both for higher earners and for low earners.
During the mid- to late 1990s, overall employment spells lasted substantially longer for those starting higher-wage than lower-wage jobs (Figure VI.3, and Tables E.3 and E.4). About 65 percent of low-wage male and female workers became nonemployed within two years after starting their jobs. In contrast, only about 45 percent of medium-wage and 39 percent of high-wage workers became nonemployed over the same period. Thus, although wage fluctuations were common for all groups of workers, unemployment was less of a problem for higher-wage workers than for low-wage ones. Again, our overall employment analysis supports these findings, because over a fixed three-and-one-half-year follow-up period, low-wage workers spent, on average, more than twice as many weeks unemployed than higher-wage workers.
|
| Source: 1996 SIPP longitudinal files using
the entry cohort sample of spells. |
What happens to low-wage workers after they leave their low-wage jobs? We have seen that during the mid- to late 1990s, about one-half of low-wage job spells ended in higher-wage employment, and about one-quarter ended in nonemployment. In this section, we examine reentry into the low-wage labor market for these workers.
a. Duration of Nonemployment Spells
During the period under investigation, about 47 percent of male low-wage workers and 57 percent of female low-wage workers exited their low-wage employment spells into nonemployment (including unemployment and leaving the labor force; Table VI.3). How long did they stay nonemployed, and what types of jobs did they find when they became reemployed?
Nonemployment spells for low-wage workers were typically short (Table VI.5). Among males in our sample, about two-thirds returned to the labor market within six months after becoming nonemployed, and 80 percent returned within a year. Reemployment rates were somewhat lower for females (51 percent found jobs within six months, and 67 percent found jobs within a year), in part reflecting the higher percentage of females who became nonemployed because they left the labor force. These relatively high reemployment rates may have been due to the strong economy faced by sample members. Nonetheless, they suggest that low-wage workers do not typically remain unemployed for a long time.
Most nonemployed low-wage workers in our sample who became reemployed returned to the low-wage labor market, and fewer entered higher-paying jobs (Table VI.5). Within 24 months after becoming nonemployed, 64 percent of males returned to low-wage jobs, compared to only 23 percent of males who found higher-paying jobs. Stated differently, more than 7 in 10 males who found jobs returned to the low-wage labor market. Similarly, more than 8 in 10 females who became reemployed returned to low-wage jobs.
| Month | Males Workers | Females Workers | ||||
|---|---|---|---|---|---|---|
| Total | Type of Reemployment | Total | Type of Reemployment | |||
| Low-Wage Job | Higher-Wage Job | Low-Wage Job | Higher-Wage Job | |||
| Cumulative Percentage of Spells Ending in Reemployment Within the Specified Number of Months | ||||||
| 1 | 18 | 14 | 4 | 13 | 11 | 2 |
| 2 | 32 | 24 | 8 | 22 | 19 | 3 |
| 3 | 42 | 31 | 10 | 29 | 25 | 4 |
| 4 | 56 | 41 | 15 | 43 | 35 | 7 |
| 5 | 61 | 45 | 16 | 47 | 39 | 8 |
| 6 | 66 | 49 | 17 | 51 | 43 | 9 |
| 7 | 69 | 51 | 17 | 55 | 46 | 9 |
| 8 | 73 | 54 | 19 | 59 | 49 | 10 |
| 9 | 74 | 55 | 19 | 61 | 51 | 10 |
| 10 | 76 | 57 | 20 | 63 | 52 | 11 |
| 11 | 78 | 58 | 20 | 65 | 54 | 11 |
| 12 | 80 | 59 | 21 | 67 | 56 | 12 |
| 13 | 81 | 60 | 21 | 69 | 57 | 12 |
| 14 | 82 | 61 | 21 | 71 | 59 | 12 |
| 15 | 83 | 61 | 21 | 72 | 60 | 12 |
| 16 | 83 | 62 | 22 | 73 | 61 | 13 |
| 17 | 84 | 62 | 22 | 74 | 62 | 13 |
| 18 | 85 | 63 | 22 | 75 | 62 | 13 |
| 19 | 85 | 63 | 22 | 76 | 63 | 13 |
| 20 | 86 | 63 | 22 | 76 | 63 | 13 |
| 21 | 86 | 64 | 22 | 77 | 64 | 13 |
| 22 | 86 | 64 | 22 | 78 | 64 | 13 |
| 23 | 86 | 64 | 23 | 78 | 65 | 14 |
| 24 | 87 | 64 | 23 | 79 | 65 | 14 |
| Source: 1996 SIPP longitudinal
files using the sample of 1,277 spells for males and 2,761 spells for females
for low-wage workers who exited their low-wage job spells into nonemployment.
Note: All figures are weighted using the longitudinal panel weight. |
||||||
b. Duration of Higher-Wage Spells
During the mid- to late 1990s, about 49 percent of low-wage employment spells for males and 36 percent of low-wage employment spells for females ended in medium-wage or high-wage employment within a four-year period (see bottom panel for each gender group in Table VI.3). In this section, we examine the rate at which workers who obtained these higher-paying jobs (1) left these jobs, (2) returned to the low-wage labor market, and (3) became nonemployed.
Our results on the duration of higher-wage employment spells show that the majority of those who obtained higher-wage jobs left these jobs within the panel period, but a significant number also remained in them (Table VI.6). Nearly 60 percent of males and females left the higher-wage labor market within one year after job start, and about 70 percent left within two years. Yet, nearly one-third stayed employed in these high-wage jobs for at least two years. Thus, we again find diversity in the labor market success of low-wage workers.
Interestingly, nearly all those who left higher-wage jobs returned to the low-wage labor market, and only a small percentage exited into nonemployment (Table VI.6). For example, more than one-half of all workers reentered the low-wage labor market within two years, whereas only about 16 percent became nonemployed over the same period. Stated another way, nearly 80 percent of those who left higher-paying jobs reentered the low-wage labor market. These results are consistent with previous findings from the overall employment analysis that many low-wage workers experienced multiple low-wage job spells during the panel period. Consequently, both exits out of and reentry into the low-wage market were common for low-wage workers during the mid- to late 1990s.
| Month | Males Workers | Females Workers | ||||||
|---|---|---|---|---|---|---|---|---|
| Total | Type of Exit | Total | Type of Exit | |||||
| Low-Wage Job | Unemployment | Left the Labor Force | Low-Wage Job | Unemployment | Left the Labor Force | |||
| Cumulative Percentage of Higher-Wage Employment Spells Ending Within the Specified Number of Months | ||||||||
| 1 | 2 | 2 | 0 | 0 | 3 | 2 | 0 | 1 |
| 2 | 6 | 4 | 1 | 1 | 7 | 4 | 1 | 2 |
| 3 | 9 | 6 | 1 | 2 | 10 | 6 | 1 | 2 |
| 4 | 36 | 30 | 3 | 4 | 37 | 31 | 2 | 4 |
| 5 | 38 | 31 | 3 | 4 | 39 | 32 | 2 | 5 |
| 6 | 39 | 32 | 3 | 4 | 41 | 33 | 3 | 5 |
| 7 | 41 | 33 | 3 | 5 | 43 | 34 | 3 | 5 |
| 8 | 50 | 40 | 4 | 5 | 52 | 43 | 3 | 6 |
| 9 | 52 | 41 | 4 | 6 | 54 | 44 | 3 | 7 |
| 10 | 52 | 42 | 5 | 6 | 55 | 44 | 3 | 7 |
| 11 | 53 | 43 | 5 | 6 | 56 | 45 | 4 | 8 |
| 12 | 58 | 46 | 5 | 6 | 60 | 48 | 4 | 8 |
| 13 | 58 | 46 | 5 | 7 | 61 | 48 | 4 | 8 |
| 14 | 59 | 47 | 6 | 7 | 61 | 48 | 4 | 8 |
| 15 | 60 | 47 | 6 | 7 | 61 | 49 | 4 | 8 |
| 16 | 64 | 50 | 7 | 7 | 65 | 51 | 4 | 9 |
| 17 | 64 | 50 | 7 | 8 | 65 | 51 | 4 | 9 |
| 18 | 64 | 50 | 7 | 8 | 66 | 52 | 5 | 10 |
| 19 | 65 | 50 | 7 | 8 | 67 | 52 | 5 | 10 |
| 20 | 66 | 51 | 7 | 8 | 69 | 54 | 5 | 10 |
| 21 | 67 | 52 | 7 | 8 | 69 | 54 | 5 | 11 |
| 22 | 67 | 52 | 7 | 8 | 69 | 54 | 5 | 11 |
| 23 | 67 | 52 | 7 | 8 | 70 | 54 | 5 | 11 |
| 24 | 70 | 54 | 8 | 8 | 71 | 55 | 5 | 11 |
| Source: 1996 SIPP longitudinal
files using the sample of 2,061 spells for males and 2,469 spells for females
for low-wage workers who exited their low-wage job spells into medium- or
higher-wage jobs. Note: All figures are weighted using the longitudinal panel weight. |
||||||||
In sum, the labor market dynamics of low-wage workers are complex. Most low-wage workers find higher-paying jobs at some point. Many, however, return to the low-wage labor market. At the same time, however, a nontrivial share of low-wage workers exit into higher-paying employment and keep these jobs for a substantial period of time. Thus, there is considerable diversity in wage progression among the low-wage worker population, although, on average, their earnings prospects improve over time.
There is substantial diversity in job and employment spell durations among low-wage workers. Is it possible to identify subgroups of workers across whom spell durations differ? Identifying these subgroups can provide policy-relevant information as to which subgroups of low-wage workers fare best in the labor market. Furthermore, the analysis can be used to check the robustness of our previous subgroup findings from the overall employment and wage progression analyses.
To keep our presentation manageable, we present subgroup findings on (1) exit rates from low-wage job spells within 12 months after job start by type of exit; and (2) cumulative exit rates from employment spells within 4, 12, and 24 months after job start. We estimated life tables, one at a time, for key subgroups of males and females defined by individual, household, and initial job characteristics.
Because our findings strongly support those presented in previous chapters, we provide less detail on the results than before. In particular, we find that the same subgroups of workers who typically had the best overall employment experiences and wage growth also had the best spell-related outcomes. The concurrence of the subgroup results is not surprising, because we
expected that subgroups of low-wage workers who experienced the most wage progression over the medium term would also be the ones most likely to exit low-wage job spells into higher-wage employment and to have the longest overall employment spells.
a. Overall Duration of Low-Wage Job Spells
Low-wage job spells are typically short across all subgroups defined by worker and initial job characteristics (last column in Tables VI.7 and VI.8). For example, during the mid- to late 1990s, 12-month cumulative exit rates for males in most subgroups ranged from 78 to 85 percent. Similarly, the cumulative exit rates for females typically ranged from 73 to 80 percent.
Nonetheless, some patterns are evident. Low-wage spells were typically longer for older than for younger workers, but as discussed in the next section, this finding masks important age differences in the states into which workers exited. More intuitively, spell durations were likely to be longer for Hispanics, those who did not attend college and those with low wages than for their counterparts. However, exit rate differences across these subgroups are not large.
b. Types of Exits from Low-Wage Job Spells
We find larger subgroup differences in exit types from low-wage job spells:
| Subgroup | 12-Month Cumulative Exit Rate for Males, by Exit Type | Total | |||
|---|---|---|---|---|---|
| Another Low-Wage Job | Higher-WagehJob with the Same Employer | Higher-Wage Job with a Different Employer | Nonemploy-ment | ||
| Overall | 20 | 21 | 12 | 29 | 81 |
| Individual and Household Characteristics | |||||
| Age (in Years) | |||||
| Younger than 20 | 25 | 13 | 7 | 43 | 88 |
| 20 to 29 | 24 | 18 | 13 | 29 | 83 |
| 30 to 39 | 15 | 25 | 14 | 25 | 79 |
| 40 to 49 | 14 | 28 | 11 | 25 | 78 |
| 50 to 59 | 13 | 27 | 12 | 22 | 75 |
| 60 or older | 10 | 19 | 8 | 25 | 62 |
| Race/Ethnicity | |||||
| White and other non-Hispanic | 21 | 22 | 13 | 26 | 83 |
| Black, non-Hispanic | 14 | 18 | 9 | 39 | 80 |
| Hispanic | 21 | 17 | 7 | 32 | 77 |
| Educational Attainment | |||||
| Less than high school/GED | 20 | 16 | 7 | 37 | 79 |
| High school/GED | 20 | 18 | 11 | 30 | 80 |
| Some college | 19 | 27 | 12 | 26 | 84 |
| College graduate or more | 19 | 27 | 20 | 19 | 84 |
| Has a Health Limitation | |||||
| Yes | 17 | 17 | 7 | 45 | 85 |
| No | 20 | 21 | 12 | 28 | 81 |
| Household Type | |||||
| Single parent with children | 19 | 15 | 8 | 39 | 81 |
| Married couple with children | 20 | 23 | 13 | 27 | 82 |
| Married couple without children | 20 | 22 | 10 | 26 | 78 |
| Other adults without children | 19 | 18 | 14 | 32 | 83 |
| Household Income as a Percentage of the Poverty Level | |||||
| 100 percent or less | 22 | 14 | 11 | 34 | 82 |
| 101 to 200 percent | 20 | 19 | 10 | 31 | 80 |
| More than 200 percent | 19 | 23 | 13 | 27 | 82 |
| Job Characteristics | |||||
| Hourly Wages | |||||
| Less than $5.00 | 20 | 18 | 13 | 30 | 81 |
| $5.00 to $5.99 | 23 | 11 | 9 | 33 | 75 |
| $6.00 to $6.99 | 22 | 18 | 11 | 30 | 81 |
| $7.00 to $7.50 | 14 | 33 | 15 | 25 | 87 |
| Hours Worked per Week | |||||
| 1 to 19 | 26 | 9 | 15 | 35 | 86 |
| 20 to 34 | 24 | 13 | 10 | 33 | 81 |
| 35 to 40 | 18 | 20 | 11 | 30 | 79 |
| More than 40 | 17 | 33 | 15 | 20 | 86 |
| Weekly Earnings | |||||
| Less than $150 | 27 | 11 | 13 | 32 | 83 |
| $150 to $299 | 20 | 18 | 11 | 30 | 79 |
| $300 to $600 | 13 | 38 | 14 | 24 | 88 |
| Owns Business | |||||
| Yes | 14 | 40 | 25 | 12 | 90 |
| No | 20 | 20 | 11 | 30 | 81 |
| Health Insurance Coverage(a) | |||||
| Yes | 18 | 26 | 14 | 24 | 82 |
| No | 21 | 17 | 10 | 33 | 81 |
| Occupation | |||||
| Professional/technical | 15 | 35 | 18 | 17 | 86 |
| Sales/retail | 23 | 27 | 10 | 20 | 81 |
| Administrative support/clerical | 17 | 20 | 11 | 30 | 79 |
| Service professions/ handlers/cleaners | 22 | 14 | 10 | 35 | 80 |
| Machine/construction/production/transportation | 17 | 25 | 13 | 27 | 83 |
| Farm/agricultural/other workers | 22 | 15 | 10 | 35 | 82 |
| Industry | |||||
| Agriculture/forestry/ fishing/hunting | 20 | 20 | 12 | 31 | 83 |
| Mining/manufacturing/construction/transportation/utilities | 18 | 25 | 13 | 28 | 83 |
| Wholesale/retail trade | 22 | 17 | 10 | 29 | 79 |
| Personal/health/other services | 19 | 18 | 11 | 32 | 81 |
| Other | 14 | 40 | 22 | 12 | 89 |
| Source: 1996 SIPP longitudinal
files using the entry cohort sample of 4,489 low-wage job spells for males.
Left-censored spells were excluded from the sample. Note: All figures are weighted using the longitudinal panel weight. a. These figures pertain to health insurance coverage from all sources, including coverage through the employer as well as from other sources. We used this variable instead of the employer-based health insurance coverage variable, because data on overall health insurance coverage is available monthly, whereas the employer-based coverage variable pertains only to jobs in progress at the time of the interview. Thus, the employer-based health insurance variable could not always be linked to the job under investigation, which led to a significant number of missing values. However, the subsets of health insurance variables overlap considerably: the source of health insurance coverage was the employer for 80 percent of those with any coverage. |
|||||
| Subgroup | 12-Month Cumulative Exit Rate for Females, by Exit Type | Total | |||
|---|---|---|---|---|---|
| Another Low-Wage Job | Higher-WagehJob with the Same Employer | Higher-Wage Job with a Different Employer | Nonemploy-ment | ||
| Overall | 22 | 14 | 8 | 33 | 76 |
| Individual and Household Characteristics | |||||
| Age (in Years) | |||||
| Younger than 20 | 29 | 4 | 4 | 44 | 82 |
| 20 to 29 | 25 | 13 | 8 | 35 | 81 |
| 30 to 39 | 21 | 15 | 7 | 31 | 75 |
| 40 to 49 | 19 | 18 | 8 | 24 | 70 |
| 50 to 59 | 15 | 17 | 9 | 26 | 67 |
| 60 or older | 14 | 14 | 1 | 33 | 62 |
| Race/Ethnicity | |||||
| White and other non-Hispanic | 23 | 15 | 8 | 30 | 76 |
| Black, non-Hispanic | 20 | 12 | 5 | 39 | 76 |
| Hispanic | 20 | 12 | 6 | 37 | 74 |
| Educational Attainment | |||||
| Less than high school/GED | 23 | 6 | 3 | 43 | 76 |
| High school/GED | 22 | 13 | 6 | 33 | 73 |
| Some college | 25 | 17 | 9 | 30 | 80 |
| College graduate or more | 19 | 22 | 14 | 24 | 80 |
| Has a Health Limitation | |||||
| Yes | 23 | 8 | 5 | 46 | 81 |
| No | 22 | 15 | 8 | 31 | 76 |
| Household Type | |||||
| Single parent with children | 24 | 11 | 6 | 37 | 78 |
| Married couple with children | 19 | 15 | 6 | 34 | 76 |
| Married couple without children | 23 | 15 | 8 | 27 | 73 |
| Other adults without children | 26 | 15 | 11 | 27 | 79 |
| Household Income as a Percentage of the Poverty Level | |||||
| 100 percent or less | 25 | 7 | 5 | 40 | 77 |
| 101 to 200 percent | 22 | 11 | 6 | 36 | 76 |
| More than 200 percent | 21 | 18 | 9 | 28 | 76 |
| Job Characteristics | |||||
| Hourly Wages | |||||
| Less than $5.00 | 24 | 10 | 7 | 37 | 78 |
| $5.00 to $5.99 | 26 | 6 | 5 | 37 | 74 |
| $6.00 to $6.99 | 22 | 13 | 7 | 31 | 73 |
| $7.00 to $7.50 | 15 | 31 | 11 | 24 | 82 |
| Hours Worked per Week | |||||
| 1 to 19 | 24 | 10 | 7 | 38 | 80 |
| 20 to 34 | 26 | 9 | 7 | 35 | 77 |
| 35 to 40 | 19 | 17 | 7 | 30 | 74 |
| More than 40 | 21 | 21 | 11 | 27 | 80 |
| Weekly Earnings | |||||
| Less than $150 | 26 | 9 | 7 | 38 | 79 |
| $150 to $299 | 22 | 14 | 7 | 31 | 73 |
| $300 to $600 | 12 | 40 | 13 | 22 | 87 |
| Owns Business | |||||
| Yes | 15 | 27 | 20 | 19 | 81 |
| No | 22 | 14 | 7 | 33 | 76 |
| Health Insurance Coverage(a) | |||||
| Yes | 18 | 19 | 10 | 28 | 74 |
| No | 26 | 10 | 5 | 37 | 78 |
| Occupation | |||||
| Professional/technical | 17 | 29 | 11 | 24 | 80 |
| Sales/retail | 25 | 10 | 7 | 35 | 78 |
| Administrative support/clerical | 19 | 23 | 10 | 26 | 77 |
| Service professions/ handlers/cleaners | 25 | 9 | 6 | 34 | 74 |
| Machine/construction/ production/transportation | 17 | 12 | 5 | 38 | 73 |
| Farm/agricultural/other workers | 21 | 7 | 5 | 49 | 82 |
| Industry | |||||
| Agriculture/forestry/ fishing/hunting | 15 | 17 | 13 | 35 | 81 |
| Mining/manufacturing/ construction/ transportation/utilities | 17 | 16 | 5 | 36 | 75 |
| Wholesale/retail trade | 26 | 9 | 6 | 35 | 77 |
| Personal/health/other services | 21 | 17 | 8 | 29 | 75 |
| Other | 18 | 31 | 17 | 10 | 77 |
| Source: 1996 SIPP longitudinal
files using the entry cohort sample of 7,401 low-wage job spells for females.
Left-censored spells were excluded from the sample. Note: All figures are weighted using the longitudinal panel weight. a. These figures pertain to health insurance coverage from all sources, including coverage through the employer as well as from other sources. We used this variable instead of the employer-based health insurance coverage variable, because data on overall health insurance coverage is available monthly, whereas the employer-based coverage variable pertains only to jobs in progress at the time of the interview. Thus, the employer-based health insurance variable could not always be linked to the job under investigation, which led to a significant number of missing values. However, the subsets of health insurance variables overlap considerably: the source of health insurance coverage was the employer for 80 percent of those with any coverage. |
|||||
c. Duration of Employment Spells
The ordering of subgroups for those with the longest to shortest employment spells (of any wage type) are similar to the ordering of subgroups discussed above. This occurs because subgroups most likely to exit into higher-wage employment were also those least likely to exit into nonemployment. Consequently, subgroups that tended to obtain higher-paying jobs also tended to have the longest employment spells. The life table results for employment spells are presented in Tables E.5 and E.6, which also show log-rank statistics to test differences in hazard rate distributions across subgroup levels. Many of the subgroup differences are statistically significant.
[ Go to Contents ]
(45) This job mobility, however, is not necessarily a negative result, because as discussed in the previous two chapters, among workers who were continuously employed during the follow-up period, those who switched jobs tended to have more positive labor market outcomes than those who remained with their initial employers. Thus, it appears that job mobility is an avenue for wage growth for some low-wage workers.
(46) The log-rank statistic compares the actual to expected monthly hazards, where the expected hazards are calculated under the null hypothesis that the monthly hazard rates are the same for each level of the subgroup. The log-rank statistic has a chi-squared distribution with the degrees of freedom equal to one less than the number of life tables being compared.
(47) The mean spell lengths pertain to those observed during the panel period, including the right-censored spells. Thus, the figures are shorter than the ultimate mean lengths of the spells. The spell durations for left-censored spells include the time spent in the spell during the prepanel period.
(48) We did not include left-censored spells when examining the durations of low-wage job and employment spells, because most of these spells ended during the panel period. Thus, the inclusion of the left-censored spells would not provide any new information.
Main Page of Report | Contents of Report
Home Pages:
Human Services Policy
(HSP)
Assistant Secretary for Planning and Evaluation
(ASPE)
U.S. Department of Health and Human Services
(HHS)
