A Job Has More Than a Wage
The focus here is the mechanism that makes two apparently similar job moves produce different outcomes: a job is valued through more than its cash wage. For the gendered U.S. job ladder, this means examining how productivity, pay, and nonwage amenities combine, and how men and women may weigh that combination differently. The useful judgment is not whether amenities “explain away” a wage gap, but whether a move raises total job value, wage value, or both.
A wage ladder is only one coordinate
A one-dimensional wage ladder assumes that every job can be ranked from low to high by pay. Under that assumption, moving upward means accepting a job with a higher wage, and a move with no wage increase is not much of an advance. The Chicago Fed paper uses a different starting point. It models jobs as combinations of productivity, wages, and amenities, so that a worker can move to a job that is better in one dimension and less favorable in another.
Productivity in this framework is not a label for a worker's innate ability. It is a property of the match between a worker and a firm: how much the match produces. A job also has an amenity level, such as more flexible scheduling or a more manageable pace. The worker receives a wage and value from the amenity. The firm pays a cost to provide the amenity and receives value from the productive match. The job's overall value depends on how these parts fit together.
That distinction matters because pay and amenities are not simply two names for the same compensation. A wage is a monetary transfer. An amenity is a feature of the work that changes what the job is like and how usable its pay may be. A predictable schedule can make a given wage easier to live with; the option to work remotely can change commuting time; a lower work pace can make a job sustainable for someone who cannot or does not want to work at a more intense pace. None of these observations implies that an amenity has a universal cash price for every worker.
The model makes the trade-off explicit. Workers value both wage and amenity, but the weight placed on the amenity can differ by gender. Firms provide amenities subject to costs, and in the baseline model they target the average amenity valuation of people searching in the economy rather than tailoring the package to a particular gender. Jobs therefore differ in the combinations they offer even before a worker decides whether to accept one.
An outside offer is evaluated against the worker's current situation and alternative options. A worker does not ask only, “Is the new wage higher?” They compare the value of the new match with the value of staying, searching again, or—when they are not employed—remaining nonemployed. The model represents this comparison with a reservation threshold: an offer must be valuable enough to accept. The threshold can change when the worker values amenities differently or has a different alternative outside the job market.
The existing job's amenity is treated as a fixed characteristic. If a worker receives an outside offer, the employer's counteroffer mainly changes the wage, not the underlying schedule or flexibility of the job. This detail matters for wage growth. A worker who values a particular amenity may accept a job with a lower-productivity match because the amenity makes the total offer attractive. Once there, the firm may have less productive surplus to share through later wage increases. The choice that improved the worker's preferred dimension can therefore limit growth in the wage dimension.
Consider two hypothetical offers. One pays more but requires a rigid schedule and a long commute. The other pays less but offers predictable hours and a workable location. A worker who places a high value on time control may choose the second offer. That choice is not irrational and it is not proof that the second job is better for everyone. It is a different ranking of the same dimensions. If analysts later compare only wages, they will see a smaller gain; if they compare only amenities, they may miss a real monetary cost.
The model's concept of a job-value ladder holds these judgments together. A move can climb the job-value ladder because the worker reaches a better combination, even if it does not climb the wage ladder. It can also climb the wage ladder while reducing total value if the new job imposes costs the wage measure does not record. The framework does not decide in advance which dimension should dominate. It explains why the answer can differ across workers and across labor-market conditions.
The diagram is deliberately more complicated than a list of wages because the relationships are the point. Workers evaluate jobs; firms provide them; jobs combine productive and nonwage features; and an offer enters an existing comparison rather than an empty ranking. The same offer can have a different value for a worker with a different reservation threshold, while the same worker can rank offers differently as employment, family demands, or alternatives change.
Preferences become sorting
The paper does not infer the amenity mechanism from theory alone. Its survey evidence shows that men and women report different trade-offs between wages and job characteristics. Among respondents who looked for work or said they might accept a job, women have reservation wages about 23% lower than men's after the paper's controls. They also prefer to work fewer hours: the controlled difference is 2.7 hours per week. At the same time, women are more likely to reject certain job changes at any offered wage. They are 11 percentage points more likely to reject a relocation, 10 points more likely to reject a doubled commute, and 7 points more likely to reject a 10% increase in required hours.
These results may look contradictory only if a lower reservation wage is treated as a general willingness to accept any job. It is not. A reservation wage concerns the monetary threshold, while the hypothetical questions reveal limits on particular nonwage conditions. A worker can require less cash pay overall and still refuse a job that destroys schedule control or imposes an unacceptable commute. The acceptance decision is multidimensional: the wage threshold and the amenity thresholds operate together.
Job satisfaction provides a related, though different, form of evidence. Wages are positively associated with satisfaction for both men and women, but the relationship is stronger for men in the paper's regressions. Usual work hours have little relationship with men's satisfaction and a negative relationship with women's. Quality-of-life and childcare benefits have a stronger relationship with women's satisfaction, while some other benefits and commute measures have more similar relationships across genders. These are conditional correlations, not direct measures of causal preferences. They nevertheless show why identical wages need not produce identical evaluations of a job.
The paper then asks whether the reported trade-offs appear in the occupations where people work. It constructs an occupation-level amenity index from several sources. Physical activity and work pace come from O*NET work-context measures. Work flexibility comes from the 2017–18 American Time Use Survey Leave Module. Teleworkability comes from the measure developed by Dingel and Neiman. The components are normalized and weighted using willingness-to-pay estimates from related research. The index is intended to make a broad, otherwise invisible dimension of occupations measurable.
That construction is useful precisely because it is limited. The index does not include every job benefit. It leaves out some benefits tied closely to monetary compensation, such as retirement benefits, stock options, and health insurance. It also measures characteristics at the occupation level and therefore misses variation among employers and jobs within the same occupation. Training opportunities and paid time off are excluded for stated reasons, including concern that some forms of training would capture future wage growth. The index is a proxy for selected amenity differences, not a complete account of everything a job offers.
Within that proxy, the sorting pattern is clear in the paper's description. Controlling for both occupation-level wages and amenities, the share of women in an occupation increases with the amenity index and decreases with occupation-level wages. The relationship is not simply that women are concentrated in all high-amenity jobs or all low-wage jobs; the joint comparison separates the two dimensions that are positively related across occupations. An occupation can be relatively attractive for its amenities even when its wage position is less favorable, and gender composition varies with that combination.
The occupation-switch results connect sorting to the ladder directly. Conditional on a worker's initial position on the amenity ladder, women experience somewhat larger amenity gains than men. Conditional on their initial occupation wage, however, they experience substantially smaller gains in occupation-level wages. The paper summarizes this as a steeper amenity ladder and a flatter wage ladder for women. In other words, a switch can represent movement toward a more valued work arrangement while yielding less monetary progress.
That conclusion still does not mean amenities fully compensate for wage differences. The index's components are incomplete, and the paper's model finds that amenities offset only part of the gender difference in job value. A benefit can be important without being equivalent to the entire wage gap. The right question is not whether a flexible schedule makes a lower wage disappear. It is how much value the worker receives from the schedule, what productivity and future bargaining opportunities accompany the job, and whether the worker had a realistic alternative.
The model also changes how we interpret observed productivity. It begins with men and women drawn from the same ex ante productivity distribution. Yet, because women place greater value on amenities in the calibration, they are more willing to accept lower-productivity jobs when those jobs offer better amenities. The model consequently generates a realized productivity gap of about 8.5%. This is an endogenous sorting result: the jobs selected differ because the attributes are valued and traded off differently. It is not evidence that women are inherently less productive.
The same logic applies to the interpretation of choice. A choice is made inside an opportunity set shaped by employers, job design, transport, schedule requirements, household responsibilities, social norms, and the value of nonemployment. The paper's controls for marital status and children have limited effects on its estimates, but that does not remove the possibility that anticipated responsibilities or institutional constraints shape the choices being observed. A lower-wage, higher-amenity destination may reflect a preference, a constraint, an adaptation to earlier conditions, or several of these at once.
A one-dimensional wage ladder hides these pathways. It treats the cash outcome as the destination and the move as an undifferentiated event. The multidimensional model asks a more precise set of questions: Which job attributes changed? Which of them did the worker value? What productivity was attached to the new match? What alternative did the worker reject? And how much room did the new employer have to raise wages when a later offer arrived?
Those questions explain how two similar switches can produce different wage ladders without making wage irrelevant. They also set up the next part of the mechanism. Preferences and job characteristics determine what an offer means, but search and employment transitions determine which offers arrive and how a worker reaches the point of accepting one. The ladder is built not only from the dimensions of a job, but from the process that brings those dimensions together.