The Long Climb Is Real—and Uneven
The U.S. evidence follows people who were observed earning at or near the minimum wage and asks how their earnings and work unfold over the decades that follow. The central result is encouraging but conditional: the average group climbs substantially, yet the path differs sharply by age and family resources. Understanding that split lets us reject the idea of a permanent minimum-wage identity without replacing it with an equally misleading promise of automatic mobility.
The average moves from the lower tail toward the middle
Consider the 1994–1998 cohort. People in the minimum-wage group—defined for this period as earning $5.50 or less per hour—begin with low annual earnings and a low relative position. Their average real annual earnings are about $13,500 in the initial period, measured in 2021 dollars. In 1999–2008, the average rises to about $29,500. In 2009–2021, it reaches about $46,500. The change is large enough to alter the basic story one might tell about these workers: the typical person in this group does not remain at the starting level.
The relative measure moves in the same direction. The group's average position in its state earnings distribution rises from about the 23rd percentile in 1994–1998 to about the 37th percentile in 1999–2008 and about the 47th percentile in 2009–2021. That is movement toward the middle, not merely a larger dollar amount in an economy where other wages may also be rising. The measure is based on annual earnings, so it reflects both pay and time worked; it is not a direct measure of hourly wage. That distinction matters because a worker can improve their annual position through a higher wage, more quarters, more hours, or a combination of them.
Participation moves as well, though it does not become identical across groups. The minimum-wage group averages 2.709 quarters employed during the initial period, 3.211 during 1999–2008, and 3.492 during 2009–2021. The group that initially earns $5.51–$7.00 averages 2.911, 3.325, and 3.524 quarters in the same periods. The initial gap narrows, but the paths do not become indistinguishable in every dimension. A career is not just a wage line; it is also the amount and continuity of work that produces the earnings.
The comparison with initially higher-paid peers makes the word convergence precise. The group that begins at $5.51–$7.00 earns about $17,500 in the initial period, $33,500 in 1999–2008, and $47,500 in 2009–2021. By the later period, the paper reports no statistically meaningful earnings difference between that group and the workers who began at or near the minimum. The group that starts above $7.00, however, remains far ahead: its average earnings are about $47,500 initially, $61,500 in 1999–2008, and $68,500 in 2009–2021.
So the minimum-wage group closes one gap without closing all gaps. That is a more informative result than either “minimum-wage workers stay poor” or “everyone catches up.” The starting wage is associated with a ranking that can change, but the later distribution still reflects where people began and what happens in between. Over the full 1999–2021 period, the minimum-wage group averages about $770,000 in cumulative LEHD earnings, compared with about $818,000 for the $5.51–$7.00 group and about $1.298 million for the group initially above $7.00. The early disadvantage does not vanish from the entire work history simply because later averages converge.
A broader employment check tells a similar, slightly more cautious story. Using national indicators for paid employment and self-employment, the researchers find evidence of work in 69.2% of years from 1999 through 2021 for the minimum-wage group, compared with 72.5% for those who initially earned $5.51–$7.00. The difference is relatively small, but it is not zero. The group generally moves upward while retaining a modest participation gap.
The timing of the climb is part of the finding. The second cohort, observed in 2005–2009, shows remarkably similar short- and medium-run career profiles. That consistency suggests the pattern is not simply an artifact of one business cycle or one narrow historical window. Still, the path is not a straight line. For older workers, the period when they are observed at the minimum wage appears as a modest dip relative to earlier and later earnings. For younger workers, the early observations must be read carefully because some have not yet spent much time in the labor market. The same graph can therefore represent a low point for one age group and an entry point for another.
There is another reason not to read the graph as a smooth biography. The average in a later period is calculated from the people who can be observed in that period, and the national employment check is needed precisely because administrative presence is an imperfect window on work. A person can leave the covered records without leaving the labor force, while another can return after a period away. The trajectory is therefore a population portrait assembled across linked records, not a claim that every named worker moves through the same sequence of jobs. Its strength is the scale and duration of the pattern; its limit is that the data cannot narrate every transition that produced it.
The climb also takes time. The initial period is followed by a first decade in which the minimum-wage group's average earnings remain below the later level and its relative position remains well below the middle. Even when the group approaches its modestly higher-paid peers, it has done so over a horizon long enough for experience, employer changes, education, and broader economic conditions to accumulate. This is why a five-year snapshot can understate eventual movement without being wrong about a worker's position at that moment. A launching point, where the evidence supports that description, is not the same as an immediate escalator.
Finally, the comparison groups show why the destination should be described in layers. The minimum-wage group closes much of its gap with workers who started only modestly higher, but it does not erase the larger gap with workers who began above $7.00. The study thus documents mobility within a distribution, not a universal equalization of outcomes. The relevant question is not whether the floor predicts the exact endpoint; it is how much of the initial ranking persists, how much changes, and for whom.
Four trajectories, not one
The strongest explanation for the aggregate climb is composition: most minimum-wage workers in the study are young, and young workers are at a stage when earnings commonly grow with experience. But age does not operate alone. The paper separates workers into four groups by age and family income, and the resulting paths show why an average can be both accurate and insufficient.
Young workers from poor families begin at about $12,500 in average real annual earnings and reach about $43,000 in 2009–2021. Young workers from richer families begin even lower, at about $11,000, but reach about $57,500. The lower starting annual earnings for young workers are largely related to fewer hours and weeks worked, not to a different minimum hourly rate. As their working lives develop, the richer-family young group rises especially sharply. Their cumulative earnings over 1999–2021 are about $952,000, compared with about $715,000 for young workers from poor families.
The older groups move less. Older workers from poor families rise from about $16,000 initially to about $29,000 in 2009–2021, while older workers from richer families rise from about $19,500 to about $34,500. Their cumulative earnings are about $504,000 and $622,000, respectively. The gap between age groups is not simply a temporary delay before the older workers catch up. The paper identifies the age division as the sharper pattern, with family income creating a durable difference within it.
That pattern changes how “upward mobility” should be evaluated. For older workers, improvement can mean recovering from a low point and returning toward an earlier position, rather than making a dramatic leap into a new part of the distribution. The paper reports that older minimum-wage workers generally have better later careers than at the moment they are observed at the minimum, but their gains are subdued compared with those of younger workers. A low wage at mid-career may therefore be evidence of a setback or constraint, while the same wage at the start of working life can be an entry point. The wage category alone cannot tell us which story is more plausible.
Relative position makes the contrast clearer. The average peak state earnings percentiles are about 59.08 for young-poor workers, 67.16 for young-richer workers, 48.25 for older-poor workers, and 53.41 for older-richer workers. These are not guaranteed destinations; they summarize the highest observed relative position in the study's period. They show that “upward mobility” can mean reaching near the middle of the distribution for one group while another remains closer to the lower half.
The age break is more consequential than a fine distinction among young workers. The paper does not find large long-run differences between those observed at ages 15–22 and those observed at ages 23–29, even among workers from poor families. The major divergence appears between workers under 30 and those aged 30–49 at the observation period. That pattern is consistent with a life-cycle process in which early work is a platform for accumulating experience or education, but it does not prove that every young minimum-wage job provides useful training or that every older worker lacks it.
The lower starting annual earnings of young workers should not be interpreted as evidence that they are already on the worst career path. People in the same hourly-wage band can work very different numbers of hours and weeks. A young person working limited hours while entering the labor market may have low annual earnings but considerable time for later work and education; an older worker earning the same hourly wage may be working more continuously and still be unable to move far above the floor. The study's combination of earnings, quarters, and relative percentiles makes this distinction visible. It also explains why an initial income ranking can reverse without implying that the initial wage was irrelevant.
Other characteristics qualify the picture further. Within the age–family-income groups, white and male workers generally show stronger future outcomes than minority and female peers, with sex differences more pronounced than race differences in the paper's comparisons. Most minimum-wage workers have a high-school education or less at the time they are observed, and the authors note that education can increase afterward, especially for the youngest workers. Those observations point toward different access to opportunity, but the study does not identify one causal channel that explains every subgroup gap.
The changing composition of the workforce also matters for how the findings travel across time. The paper reports that young workers make up 74% of its combined minimum-wage sample. In a later comparison discussed by the authors, the share of 15–22-year-olds among minimum-wage workers falls from 58% in the 1993 SIPP to 45% in the 2022 Current Population Survey, while the shares of workers aged 23–29 and 30–49 rise. If the group observed at the floor contains relatively more adults who are already further into their working lives, its average future growth should be less steep even if the underlying labor market has not become more deterministic.
Composition is not a statistical nuisance to be averaged away. It is part of the economic meaning of the minimum-wage population. A headline that says the group rose from the lower tail toward the middle partly describes what happens when many people enter at a young age and then accumulate experience. A headline about limited mobility among older workers describes a different exposure to the same wage floor. Comparing the two without naming the age mix can make either group seem more representative than it is.
This is the key lifecycle lesson: a minimum-wage observation has a different meaning at different ages. For a young person, it may be an early position before experience, education, and employer changes accumulate. For an older person, it may mark a period of constrained opportunity or a dip from a higher prior level. Family resources can affect the ability to turn early work into later options, while gender and race can shape the opportunities available within the same age and income category. None of these factors makes a path inevitable, but ignoring them makes the average misleading.
The trajectory evidence is therefore powerful in a specific way. It shows that a low starting wage does not fix a worker's later position and that many young workers move substantially upward. It also shows that the same conclusion cannot be extended uniformly to older workers, workers from poorer families, or the current workforce without attention to composition. The data describe a population moving through time; they do not tell us that the initial job caused the movement.
That final distinction matters before any policy is evaluated. Age, experience, family resources, education, selection into work, and regional growth can all help produce an upward career profile. A policy change must be judged against those forces, not credited with every later improvement. The next question is accordingly narrower and more demanding: when the wage floor itself rises, does it change the worker's path beyond the movement already associated with age and place?