we-solve (beta)

Econ-444 · economics · Economics

Wages versus job amenities

LLM status Partially solved unverified

Statement

Original subject group: Labor markets and work

Audit classification: Background; proposed extension. Full review checked. It explains identification and confounding; its conclusion does not explicitly pose this wage-equivalent inequality problem as open.

Definitions and precise research target

Complete editorial problem. Fix a worker population and an observable set of jobs. Job $j$ supplies wage $w_j$ and amenity vector $a_j$, including scheduling, security, location, and workplace conditions. Worker $i$ has utility $u_i(w,a)$, strictly increasing in wage, and a feasible consideration set $C_i$. For a declared common amenity bundle $a_0$, define wage-equivalent job value $e_{ij}$ by $u_i(e_{ij},a_0)=u_i(w_j,a_j)$, when this inverse exists. Let $j(i)$ denote actual employment. The target is the distribution of $e_{i,j(i)}$, together with its inequality relative to the distribution of observed wages. Identify or bound the required preferences using credible offer variation, stated choices validated against behavior, and information about $C_i$. Separate compensation for amenities from firm productivity, rent sharing, and endogenous matching. Report how conclusions depend on the reference bundle and heterogeneous preferences; interpersonal utility comparisons need additional declared normalization. A policy counterfactual equalizing amenities must additionally model wage and job reallocation. The proposed measurement task is to quantify these adjustments reliably in a specified population, not to infer their sign from competitive hedonic regressions alone.

Definitions and admissible scope. Let the wage domain be an interval $\mathcal W\subseteq\mathbb R$ and amenity space $\mathcal A$. Assume $u_i(\cdot,a_0)$ is continuous and strictly increasing. Require $u_i(w_j,a_j)$ to lie in its range; this guarantees a unique wage equivalent $e_{ij}$, which need not be positive. If $E|e_{i,j(i)}|<\infty$, define its distribution under a fixed worker measure. Use a variance comparison only under finite second moments; use Gini or top-share comparisons only when their denominators and domain conditions hold. Job choices additionally depend on consideration sets, offer probabilities, match-specific utility, and tie-breaking. Interpersonal comparison of money equivalents is conditional on the same currency, price base, and reference amenity bundle, not a utility normalization inferred from a hedonic regression. The baseline distributional target is restricted to a fixed employed cohort, with positive wages and $j(i)$ defined for each included worker. If nonemployed persons are included, specify a separate outside-option bundle before computing its money equivalent. Choose the inequality functional as an input with a stated domain; for example, $\operatorname{Var}(e)$ requires finite second moments. The data input includes offered and accepted jobs and experimentally or observationally identified choice probabilities. The admissible preference/consideration-set family and measurement restrictions determine which value distributions are compatible with those probabilities. Return that identified set rather than claiming that wage data alone recover $e$.

Variants and assumption changes

  1. Experimental offers (editorial specialization informed by the review): randomize wages and one amenity over common offer support. Identify willingness to accept for that amenity under stable preferences and known offer sets; report only bounds outside the experimental wage range.
  2. Equilibrium equalization (editorial): set $a_j=a_0$ for all employers and recompute wages, costs, participation, and sorting in a declared model. Compare inequality after adjustment with the baseline money-equivalent distribution; the two are different objects.

References and source audit

Support and access. Full review checked. It explains identification and confounding; its conclusion does not explicitly pose this wage-equivalent inequality problem as open. Locator: Sections 8.2 and 9, printed pp. 60--64 and 75--76.
  1. Alexandre Mas. 2025. Non-Wage Amenities. Section 8.2, printed p. 63; conclusion, printed pp. 75–76. https://www.rfberlin.com/wp-content/uploads/2025/08/25051.pdf. DOI: https://doi.org/10.3386/w33643.

Integrated refinement: ECON-073

Valuing security and other nonwage attributes of informal jobs

This refinement adopts the additional source conventions (audited:conventions), subject to its local restrictions. Informal-job security and constrained offer sets. Specialize amenities to $(w,h,r,b,a)$: wage, hours, separation risk, benefits and autonomy. For a security improvement $r_0\to r_1$, define compensation by $U_i(w-CV_i,h,r_1,b,a)=U_i(w,h,r_0,b,a)$ when a unique solution exists. Identify the distribution of this and other attribute valuations using specified preference, offer and measurement models. Fix currency, horizon, reference bundle and baseline worker population; specify utility monotonicity, the income range and outside options. Accepted informal/formal jobs can reflect constrained or selected offers rather than unconstrained preferences. Separate wage changes, benefits and security; validated offer variation is needed to attribute a valuation to one attribute.

Supplied source audit for ECON-073

Local [S1], [S2], etc. in this audit and the references immediately below belong to ECON-073, independently of the original entry's reference numbering. [S1] and [S2] support valuation of nonwage work conditions. The compensation equation is editorial and is conditional on the stated bundle and offer constraints.

References for integrated refinement ECON-073

[S1] Gabriel Ulyssea, Matteo Bobba, Lucie Gadenne, and Mariaflavia Harari (2025). Informality. VoxDevLit 6(2), April 2025.

Verified locator / source role: Primary publisher page checked for review identity, authors, version and background. The specific published agenda is corroborated in [S2]; the exact formal target is editorial.

https://voxdev.org/voxdevlit/informality

[S2] Oliver Hanney / VoxDev (living compilation). Evidence gaps: Key unanswered questions in development economics. Accessed 8 October 2026. The page currently covers 20 topics; the "16-topics URL is a historical slug.

Verified locator / source role: Topic heading: Informality. Supports the broad research agenda; this is not a verbatim quotation or an attribution of the displayed mathematical target.

https://voxdev.org/topic/evidence-gaps-key-unanswered-questions-across-16-topics-development-economics

Common conventions: Source mathematical conventions

These conventions apply to each entry unless explicitly replaced.

Models and identification. A primitive model $m$ specifies all probability laws, feasible choices, information, equilibrium and measurement rules used by its target. The admissible class $\mathcal M$ is fixed independently of outcomes. If $P_O(m)$ is its observable probability law and $\tau(m)$ the target, its identified set at observable law $P$ is \[ \mathcal I_\tau(P)=\{\tau(m):m\in\mathcal M,\ P_O(m)=P\}. \] Point identification means that this set is a singleton. An empty set signals incompatibility of data and assumptions. Coordinate bounds need not describe the joint set. Bounds are sharp only if every retained target is attainable or arbitrarily approachable by compatible models. Finite bounds require explicit restrictions on unobserved quantities; unrestricted confounding, hidden wealth, extrapolation or arbitrary equilibrium selection can yield uninformative sets.

Causal interventions. A potential outcome $Y_i(p)$ is the outcome under a complete policy regime $p$, including timing, financing, information and market spillovers. The target population and units are fixed before treatment. With interference, use the complete assignment vector or a justified exposure mapping; individual no-interference notation alone cannot identify an economy-wide intervention. A difference of mean potential outcomes does not require the cross-world joint distribution. Conditioning on post-treatment migration, survival, enrollment or employment changes the estimand and needs separate justification.

Statistical statements. An estimator, confidence set, forecast or test specifies a sampling law, model class and loss/coverage criterion. Uniform $(1-\alpha)$-coverage means $\inf_{m\in\mathcal M}\Pr_m\{\tau(m)\in C_n\}\geq1-\alpha$, or an explicitly asymptotic version. The whole parameter space trivially has coverage, so informativeness requires a stated width, risk or power objective. A forecasting improvement is relative to a named benchmark, information set and evaluation population.

Equilibrium and optimization. An equilibrium comprises feasible plans, optimality, beliefs where needed, budget constraints and all market/resource clearing. The primitives or a stated benchmark define each correspondence $\mathcal E(p;m)$. Nonemptiness is assumed or proved, never inferred just from compact policy sets. A selected-equilibrium target uses a declared selection rule; a robust target quantifies over all permitted equilibria. Use suprema or infima unless attainment is established. An $\varepsilon$-optimal policy is feasible and within $\varepsilon$ of the finite optimum value. Report an empty feasible set separately.

Welfare and accounting. Normative utility, interpersonal normalization, social weights, numeraire, discounting and the use of public receipts are primitives. Behavioral utility need not coincide with the welfare criterion. Household welfare includes ownership income and rents once; adding profits again to compensating variation generally double-counts them. A monetary equivalent is defined by an explicit transfer and price convention, with monotonicity and range conditions. Average effects, mechanism contributions, transfers and welfare losses are different targets. Infinite sums require convergence and dynamic feasible plans require terminal or no-Ponzi conditions.

Source versions. A working-paper date, revision date and journal publication year are distinguished. A DOI for a journal version is not automatically the DOI of an earlier manuscript. Locators refer to the stated version and printed pages where specified. A metadata-only check does not verify a claim about a full-text conclusion. Every entry's source subsection narrows the provenance claim accordingly.

Common conventions: Additional-source status and mathematical conventions

Of the 100 supplied ECON entries, 65 distinct problems appear here; 32 close matches refine earlier entries, and three duplicate questions map to existing entries. Appendix app:audited-crosswalk lists every destination. The attachment's P/R/S/U distinctions and source audits are retained. Its precise mathematical formalizations are editorial unless the individual audit says otherwise. Candidate subsidy targets ECON-004--006 retain their U flags; the resolved approval-core question ECON-007 updates OP-0202 above. The following supplied conventions govern both this part and the attributed ECON refinements elsewhere in the catalogue, subject to local restrictions.

Parameterized questions. A problem family lists inputs that must be fixed before an instance is evaluated: population, horizon, policies, information, data law, model class and welfare weights. These are required choices, not quantities the citations are assumed to specify. Undefined applications should not be treated as identified estimands.

Indices and integrability. Sets of agents, firms and regions are finite unless a population measure is explicitly used. $i,j$ index units, $r$ regions, $t$ dates and $h$ horizons. $\mathbb E$ and $\Pr$ refer to the declared population and law; $\mathbf1\{A\}$ is an indicator. Every displayed expectation and discounted sum needs existence in the stated domain. Infinite horizons use a discount in $(0,1)$ and a convergence condition; finite horizons may use discount one.

Optimization and existence. A displayed $\max$, $\min$, or $\arg\max$ is conditional on attainment. For finite feasible menus this is immediate; general spaces need continuity and compactness/coercivity or another existence argument. Without attainment, the target is the supremum/infimum and arbitrarily near-optimal feasible choices. $\inf\varnothing=+\infty$. Empty feasibility and an unidentified target are different issues.

Potential outcomes. $Y_i(z)$ is the outcome under a specified intervention, not the observed conditional mean among people choosing $z$. In binary comparisons $\Delta Y_i=Y_i(1)-Y_i(0)$. Specify assignment, treatment versions, compliance, information and observation horizon. Interference is allowed under a full policy vector; compressed exposure notation needs a justified mapping. No interference, consistency and overlap are substantive assumptions, not automatic consequences of notation.

Populations and observation. Fix baseline eligibility and population weights. Conditioning on post-policy employment, survival, relocation or program uptake can change the population being compared. Death is not ordinary loss to follow-up; outcomes truncated by death need a composite convention, joint survival target, or explicitly defined survivor stratum. Fertility-changing interventions need a population functional rather than an unexplained fixed descendant cohort. Measurement and missingness models must be supplied.

Identification. A structural model $m\in\mathcal M$ implies observable law $P_m$ and target $\theta(m)$. Identification means \[ P_{m_1}=P_{m_2}\ \Longrightarrow\ \theta(m_1)=\theta(m_2) \quad\text{for all }m_1,m_2\in\mathcal M . \] Without identification, the target is \[ \Theta(P;\mathcal M)=\{\theta(m):m\in\mathcal M,\ P_m=P\}. \] Writing $\theta(P)$ before establishing identification can hide incompatible latent models. Confidence sets additionally need a sampling law, confidence level, asymptotic sequence and uniformity class. Point identification, coverage and informative precision are separate properties.

Mediation. Total effects of randomized assignment are distinct from cross-world natural indirect effects. Belief/effort-channel effects require a meaningful mediator intervention and additional measurement, exclusion or mediation assumptions. Randomization of the initial policy alone does not supply those assumptions. Report bounds or interventional alternatives when the stronger target is unsupported.

Equilibrium. A counterfactual $W(z)$, $p(z)$, or $\mu_t^z$ requires individual optimization, feasibility, government/resource budgets, market clearing and state-transition laws. State equilibrium existence and selection, or report ranges over compatible equilibria. Initial-state integration includes subsequent endogenous transitions; current-state integration must use the policy-dependent distribution. Reduced-form invariance after policy is not assumed.

Welfare accounts. Scores, utility, health, dollars and ecological quantities require explicit conversion before addition. Fees, taxes, subsidies, wages and extortion payments are transfers between parties; distributional weights can make them welfare relevant, but their face value is not automatically a resource loss. State ownership, geographic boundaries, foreign-income weights and fiscal finance. Do not add profits to household welfare already including dividends, or count land capitalization and the underlying benefit twice. A benefit-cost expression must distinguish gross benefits from net welfare and subtract every real cost exactly once.

Derivatives and risk. Log derivatives require positive arguments; local responses require admissible perturbations and specified equilibrium branches. Differentiation under expectations needs regularity/domination, and boundaries may allow only one-sided derivatives. Risk uses a specified law; ambiguity uses a declared nonempty law set on a common history space. Adaptive policies must be nonanticipative. A robust criterion is an editorial choice unless attributed explicitly.

Scope overlaps. Allocation, collective choice, contracts, household bargaining and coordinated policy overlap economics with optimization and game theory. They are retained because they appeared in the original catalogue; no claim of a strictly game-theory-free selection is made.

Statement licence: CC-BY-4.0 · Source file

Contributions 1

Econ-444-A1 · Partial result · LLM-generated · gpt 6 astra ultra ·

GPT 6 Astra Ultra attempt on Econ-444

Offer experiments and sharp money-equivalent inequality sets

Status: Partial; the full catalogue problem remains unresolved.

Abstract

In a quasi-linear one-amenity model, randomized offers identify willingness-to-pay distributions only over the experimental wage range and only for workers with the stated offer set. Finite individual choice records generate sharp intervals for wage-equivalent job values. Their mean and variance identified sets can be computed exactly. Examples show that amenity adjustment can either increase or decrease inequality, and that changing the reference bundle can change the answer when preferences differ.

A deliberately restricted preference model

The catalogue target is the distribution of wage equivalents in a fixed employed cohort, not the wage distribution after all employers are forced to offer the same amenities. Fix $n$ workers with weights $q_i>0$, $\sum_iq_i=1$. Their observed jobs have wage $w_i$ and scalar amenity $a_i$. Preferences are \[ u_i(w,a)=w+\theta_i a, \tag{1} \] with a stable but initially unknown $\theta_i$. Wage has the same units for everyone. The amenity can represent one standardized scheduling or security improvement; a composite interpretation needs stronger assumptions. We allow negative $\theta_i$ unless a stated domain rules them out.

For a common reference $a_0$, equation (1) gives the exact wage equivalent \[ e_i=w_i+\theta_i(a_i-a_0). \tag{2} \] Strict wage monotonicity guarantees uniqueness over an unrestricted real wage domain. If the permitted wage domain is bounded, (2) is usable only when it lies inside that domain. The variance functional below permits negative money equivalents, whereas a conventional Gini with a positive-mean interpretation needs additional restrictions.

Each experimental offer compares $(w,0)$ with $(w-D,1)$, and both offers are known and genuinely feasible. Choice is deterministic given preferences, with a prespecified tie rule. Assignment of $D$ is independent of $\theta_i$ in the sampled population. There are no offer-specific shocks, liquidity restrictions, or treatment-induced preference changes in this benchmark.

What a randomized wage discount identifies

Proposition: The worker chooses the amenity offer exactly when $\theta_i\ge D$, with equality governed by the tie rule. Hence offer-choice probabilities identify the survival function of $\theta$ on the randomized support of $D$. They do not identify its distribution outside that support.
Proof: The utility difference is $(w-D+\theta_i)-w=\theta_i-D$. Independence of the discount from preferences makes the conditional choice probability equal to the population mass above the threshold, including any tied mass according to the rule. Atoms can be recovered using the appropriate one-sided limits where those limits are supported. To prove the support qualification, take two distributions agreeing on the entire experimental discount interval, including the total mass in each external tail, but locate their lower-tail or upper-tail masses at different unsupported values. Their choices coincide for every tested discount while their tail valuations differ. ∎
One cross-sectional experiment identifies a marginal distribution of $\theta$, not each worker's realization. Applying it to (2) also requires its dependence on baseline wage and amenity. For example, randomization within recorded $(w_i,a_i)$ cells identifies conditional valuation distributions in those cells, subject to overlap. An experiment on a separate convenience sample does not identify the covariance of wages and preferences in the target cohort merely because its marginal wage distribution looks similar.

Known consideration sets are equally important. If the amenity offer is unavailable to some workers, rejection need not mean $\theta_i<D$. Without a model for availability, even a randomized displayed offer can mix preference responses with constraints.

Individual intervals and exact finite-cohort bounds

Consider a richer panel of validated binary offers for each worker. An accepted amenity offer at discount $D$ gives a lower inequality for $\theta_i$; a rejected offer gives an upper inequality. Combine all these observations with a declared finite support restriction $\theta_i\in[\underline\theta_i,\overline\theta_i]$. Write the resulting closed interval as $[\ell_i,h_i]$. If any such interval is empty, the maintained preference model is rejected; the following bounds concern the nonempty case. If strict revealed inequalities occur, the same endpoints describe the closure of the identified set, with infima and suprema approached rather than necessarily attained.

Let $b_i=a_i-a_0$ and define \[ L_i=w_i+\min\{b_i\ell_i,b_ih_i\},\qquad U_i=w_i+\max\{b_i\ell_i,b_ih_i\}. \] Assume the admissible preference family imposes no further cross-worker restrictions.

Theorem: The sharp closed identified set for the vector of equivalents is the box $\prod_i[L_i,U_i]$. Its mean ranges over \[ \left[\sum_iq_iL_i,\ \sum_iq_iU_i\right]. \tag{3} \] For variance $V(e)=\sum_iq_ie_i^2-(\sum_iq_ie_i)^2$, the sharp lower and upper endpoints are \[ V_{\min}=\min_{e_i\in[L_i,U_i]}V(e),\qquad V_{\max}=\max_{e_i\in\{L_i,U_i\}}V(e). \tag{4} \] The full variance identified set is the interval $[V_{\min},V_{\max}]$.
Proof: For $b_i\ne0$, the affine map in (2) maps the preference interval onto $[L_i,U_i]$; for $b_i=0$ the equivalent is fixed at $w_i$. Independent admissibility of the preferences makes every point of the box compatible. Linear optimization proves (3).

The matrix of the quadratic variance is $\operatorname{diag}(q)-qq^\top$. For any vector $v$, its quadratic form is $\sum_iq_iv_i^2-(\sum_iq_iv_i)^2\ge0$ by the weighted square inequality, so minimizing variance is a convex quadratic problem on a compact box. A maximum exists as well. Holding the other coordinates fixed, variance is convex in each $e_i$, with quadratic coefficient $q_i(1-q_i)\ge0$. Moving one coordinate to an endpoint can therefore weakly increase its value. Applying this successively to all coordinates shows that a maximizing vertex exists, establishing the upper formula in (4). Finally the box is connected and variance is continuous, so its image is a connected compact subset of the real line: exactly the interval between its extrema. ∎

The theorem is computationally explicit: minimum variance is a bounded convex program, while exhaustive evaluation of the $2^n$ vertices gives an exact upper endpoint. This is a finite characterization, not a claim of a polynomial-time method for the upper problem. Dependence restrictions across workers would change the feasible set and may tighten both endpoints.

For two equally weighted workers, $V(e)=(e_1-e_2)^2/4$. If their equivalent intervals are $[1,3]$ and $[2,4]$, equality is feasible and $V_{\min}=0$. The largest distance is $3$, so $V_{\max}=9/4$. Every intermediate value is attainable. Reporting just the variance obtained by putting both workers at their interval midpoints would give an unjustified point estimate.

Inequality can move in either direction

Take equal weights, observed wages $(1,3)$, common preference $\theta=2$, and reference $a_0=0$. Observed wage variance is $1$. If amenities are $(1,0)$, equivalents are $(3,3)$ and variance is zero. If amenities are $(0,1)$, equivalents are $(1,5)$ and variance is $4$. Both cases obey the same preference model and have the same wage distribution. Wage inequality alone therefore does not determine the sign of the amenity adjustment.

When $\theta_i$ differs across workers, even the reference bundle matters. Let $e_i(0)=w_i+\theta_i a_i$. Then \[ {\operatorname{Var}}_q(e(a_0)) ={\operatorname{Var}}_q(e(0))-2a_0{\operatorname{Cov}}_q(e(0),\theta) +a_0^2{\operatorname{Var}}_q(\theta). \tag{5} \] This follows by expanding the variance of $e_i(0)-a_0\theta_i$. With common preferences the reference change subtracts a constant and leaves variance unchanged; heterogeneous preferences remove that invariance. The reference bundle is therefore part of the estimand, not an innocuous plotting choice.

The baseline accounting in (2) also differs from a policy that changes every actual amenity to $a_0$. Such a policy may change employer costs, wages, employment, and sorting. Holding $w_i$ fixed in that counterfactual would require a separate equilibrium argument.

Unresolved components

The sharp results apply to stable quasi-linear preferences, validated offer sets, finite valuation support, and an individual panel or correctly linked conditional experiment. The full research question allows nonlinear utility, many amenities, unemployment, preference shocks, constrained search, and endogenous matching. Those features are not estimated here. In particular, lifetime job security may require an intertemporal expected-utility model rather than the linear amenity index in (1).

The offer-threshold argument and finite-box optimization are standard tools, proved here to expose exactly which inequality claims the data can support. There is no claim that the sign of wage-equivalent inequality has been resolved for any real population.

Completion Estimate

60%

This is a post hoc, heuristic estimate for the full catalogue target. The attempt identifies supported offer-threshold valuations and proves sharp finite-cohort wage-equivalent mean and variance sets, including reference-bundle sensitivity. Nonlinear multi-amenity utility, intertemporal security, constrained search, measurement uncertainty, and equilibrium job reallocation still exceed the restricted preference model.

References

  • [mas] A. Mas (2025). Non-Wage Amenities. RFBerlin discussion paper, sections on identification and the conclusion. The review motivates caution about wage regressions and job attributes; the restricted box characterization here is self-contained. https://www.rfberlin.com/wp-content/uploads/2025/08/25051.pdf.

Go to contribution →