Four distributive principles organize the normative debate. Pereira et al.( 2) reviewed them for transport and showed that egalitarianism, sufficientarianism, prioritarianism and utilitarianism each yield a different allocation rule when applied to accessibility. Egalitarianism demands equal accessibility for all. Prioritarianism weights gains toward those who start worse off. Sufficientarianism guarantees a minimum threshold. Utilitarianism maximizes total benefit regardless of its distribution. Pereira( 19) restated these distinctions for a transport audience and noted that the four principles cannot all be satisfied at once. Randal et al.( 20) reached the same conclusion and argued that the choice among principles is a political decision that appraisal cannot outsource to technique. van Wee and Roeser( 7) made the point against cost-benefit analysis specifically, whose utilitarian aggregation of travel time savings hides distributional judgments that egalitarian frameworks would make visible.

The differences among the principles are consequential in practice. A fare cut financed by general taxation satisfies utilitarians if total welfare rises, satisfies prioritarian readers only if the gain reaches the transport poor, and satisfies egalitarians only if it narrows the gap between groups. Jeekel and Martens( 21) observed that housing, health care and education policy institutionalized such distinctions decades ago, whereas transport still treats equity as an afterthought in appraisal. The comparison suggests that the problem is not a missing principle but a missing institutional commitment.

The practical choice between principles can be illustrated with a single target. A sufficiency standard for job access asks how many residents can reach a given number of jobs within a fixed time, whereas a capabilities standard asks whether residents can convert that reach into employment given their skills, schedules and care duties.( 22) The two questions require different data and produce different beneficiaries, which is why the choice cannot be postponed until measurement begins.( 20)

The genealogy of the principles matters for their authority. The transport field imported them from political philosophy, mostly through review articles rather than through original philosophical work, which means the principles arrive pre-digested and occasionally simplified.( 2) Vecchio and Martens( 22) warned that the capabilities approach in particular is invoked without its full machinery of conversion factors and agency, reducing a demanding framework to a label. Normative depth therefore varies across studies, and depth, in turn, predicts whether a study names its principle.( 3)

Prioritarianism and sufficientarianism overlap enough to confuse operational choice. Both direct attention to the worst off, and both can be satisfied by a threshold, but they diverge on whether gains above the threshold matter for equity.( 20) A prioritarian index keeps counting improvements for the worst off even after a minimum is met, whereas a sufficientarian index stops at the threshold. Empirical work that distinguishes the two is scarce, which is one reason why studies that claim a sufficientarian reading can be re-read as prioritarian.( 23)

Sufficientarianism has gained ground because it replaces the demanding comparisons of strict egalitarianism with a minimum standard. Ryan et al.( 23) asked what a sufficient level of accessibility would be and why no transport agency has defined one, while Rode et al.( 24) traced a sufficiency turn in urban transport policy. The principle is attractive in practice, because a threshold of access to jobs, health and education can be written into plans and audited. Its weakness is that no consensus exists on where the threshold lies, and studies that propose one remain scarce.( 23)

The capabilities approach offers a different grammar. Instead of distributing a resource, it asks what people can actually do with transport, and therefore foregrounds conversion factors such as income, disability, care responsibilities and fear.( 22) Luz and Portugal( 25) re-read transport-related social exclusion through that lens, and Azmoodeh et al.( 26) identified the factors that shape transport capability and argued that equal access does not guarantee equal capability. Applied to older adults, Deka( 27) interpreted trip deprivation through the capability approach rather than through service supply, and Cooke et al.( 28) warned that physical proximity is not access for pedestrians who cannot traverse the route.

The divergence among these frameworks is real but productive. They agree that distribution matters and that accessibility is its object; they disagree on the unit, resource, threshold or capability, and on how much inequality is tolerable. Yet none of them specifies an index. That silence is the hinge on which the measurement problem turns.( 3,4)

The second axis documents the consolidation of accessibility as the currency of justice. The position originates in the claim that the transport system's distributive good is access to destinations rather than movement itself, an argument that Pereira et al.( 2) developed into a full mapping of distributive principles onto accessibility. van Wee( 4) formalized the link in a framework that separates accessibility, its determinants and its equity evaluation, and Jeekel and Martens( 21) showed that transport lags behind other social policy domains in institutionalizing that separation.

The clearest consequence concerns appraisal. Martens and Di Ciommo( 6) demonstrated that travel time savings, the standard benefit measure, reward those who already travel and can therefore widen accessibility gaps even when a project appears beneficial. Di Ciommo and Shiftan( 29) showed that accessibility-based evaluation reverses conventional project rankings. van Wee and Roeser( 7) supplied the ethical grounding for that reversal. The convergence is strong: appraisal without accessibility is blind to distribution.

The fifteen-minute city debate has brought this axis into policy. Vale et al.( 11) measured pedestrian accessibility to everyday destinations across European cities and documented wide inequality between centers and peripheries, while Knap et al.( 12) built a cycling accessibility metric for Utrecht that connects proximity to socioeconomic inequality. These studies share the accessibility-as-currency premise, yet they stop short of naming the distributive principle that would turn a proximity gap into a policy obligation.( 4)

The persistence of mobility in practice is itself a finding. Despite the conceptual consensus, speed and travel time remain the metrics that planning agencies report, because they are cheap to compute and legible to elected officials.( 6) Di Ciommo and Shiftan( 29) observed that the shift from mobility to accessibility is therefore incomplete at the operational level, even where it is complete at the conceptual level. The gap between the two levels mirrors, in miniature, the larger gap between theory and measurement that this review traces.

The boundary between accessibility and mobility is substantive rather than terminological. Mobility measures count trips and speeds, which reward travel itself, whereas accessibility measures count reachable opportunities, which reward the ability to satisfy needs without necessarily traveling.( 4) Pereira( 19) illustrated the distinction with teleworking: a service that removes a commute leaves mobility unchanged but raises the accessibility of work. Policy built on mobility therefore incentivizes movement, while policy built on accessibility incentivizes proximity and connectivity.

The participation turn sharpens the accessibility claim. Allen and Farber (30) reason that accessibility is only instrumentally valuable, because what justice ultimately protects is the ability to take part in work, care and civic life. On that view, an accessibility index is a proxy whose validity depends on how tightly accessibility predicts participation. The available evidence supports the proxy but not its sufficiency: Luz and Portugal (31) found that accessibility reduces social exclusion in São Paulo, while Ward et al.(32) found that disadvantage reshapes the activity set independently of measured access. The two findings are compatible only if accessibility and participation are treated as distinct, linked constructs.

The social exclusion tradition provides the empirical counterpart of this normative claim. Recent work confirms that low accessibility compounds disadvantage across life domains. Ward et al.( 32) found that transportation disadvantage narrows the set of activities people even consider, and Luz and Portugal( 31) estimated in São Paulo that higher accessibility reduces the probability of social exclusion. Allen and Farber( 30) argued that participation in activities, not accessibility alone, is the better outcome, a position that complicates the consensus without overturning it. Lunke( 33) showed in Oslo that modal accessibility disparities map onto transport poverty, and Venter et al.( 34) reviewed bus rapid transit evidence to find that infrastructure improvements do not automatically reach low-income riders.

The axis also records a persistent boundary debate. Mattioli et al.( 35) linked affordability to exclusion in the United Kingdom, and Bueno Rezende de Castro et al.( 36) showed that transport poverty leads to social exclusion in a low-income community. Latin American case studies add that exclusion persists even under subsidized fares.(34,37) The divergence is geographic and conceptual, because exclusion is mediated by fares, service and land use in ways a single accessibility index cannot capture.

The third axis concerns how accessibility inequality is quantified. Distributional indices adapted from income research dominate. Raza et al.( 38) applied Lorenz curves and the Gini coefficient to time-dependent public transport accessibility, and the approach has since been extended to food store access during the pandemic,( 39) rail network access( 40) and 15-minute pedestrian accessibility across European cities.( 11) The Gini coefficient summarizes the whole distribution in one value and permits comparison across cities and years; its cost is that it hides where in the distribution inequality sits. The Theil index and the Palma ratio answer that weakness by decomposing inequality between and within groups, though they appear in a smaller number of studies.( 38)

Vertical and horizontal equity provide the other dominant frame. Karner( 10) operationalized vertical equity through needs-based gaps between disadvantaged and reference groups, and Nahmias-Biran et al.( 41) measured the distributional effects of a fare change in Haifa with Lorenz-derived indicators. Yeganeh et al.( 42) applied a social equity analysis to job accessibility across the United States public transportation system and found that minority and low-income groups face systematically lower access to jobs. Camporeale et al.( 43) embedded horizontal and vertical equity directly in a transit design model, and Ruiz-Pérez et al.( 44) showed that the chosen zoning changes the horizontal equity verdict, a modifiable areal unit problem. Lewis et al.( 3) surveyed the norms actually used and found them implicit and inconsistent across studies.

A third finding is that measurement choices materially change the result. Kapatsila et al.( 45) compared cumulative and gravity-based accessibility in eight Canadian cities and found the two metrics rank neighborhoods differently, which in turn changes which areas appear disadvantaged. Pereira (46) showed that the travel time threshold chosen for a scenario analysis changes the sign of the equity effect, and Stępniak et al.( 13) demonstrated that temporal resolution alters measured public transport accessibility. Dianin et al.( 47) isolated the transport system's contribution from individual constraints, a step most equity indices skip.

The scale at which inequality is measured remains a further source of divergence. Place-based indices, computed for zones or neighborhoods, cannot see differences within households, while person-based indices require travel survey or mobile data that are costly and inaccessible to most researchers.( 47) Tiznado-Aitken et al.( 15) showed in Santiago that accounting for level of service and competition for opportunities changes the measured distribution of public transport accessibility. The same city can therefore appear equitable at one scale and sharply unequal at another.( 44)

Data availability shapes which measures are feasible and therefore which norms get expressed. Distributional indices require only an accessibility score per zone, which explains their dominance; needs-gap and capability measures require group-specific thresholds or individual conversion factors, which are harder to obtain.( 48) Stępniak et al.( 13) showed that even the accessibility score depends on timetable resolution, so the data question is not settled by using public transport schedules. The measurement literature thus faces a double constraint: the norm it might prefer and the data it can afford.

Beneath the index choice lies an older distinction between equality of opportunity and equality of outcome. A Gini coefficient of accessibility summarizes an outcome distribution, whereas a needs-gap or vertical equity ratio approximates an opportunity comparison between groups.( 10) The two framings answer different questions and therefore lead to different interventions. Studies do not state which framing they adopt, and the silence reproduces the same ambiguity as the missing normative principle.( 3)

The choice between place-based and person-based equity also has political consequences. Place-based indices identify deprived neighborhoods and suit area-based programs, whereas person-based indices identify deprived individuals and suit targeted subsidies.( 47) The two unit types can disagree about where the problem lies, and studies that report only one silently foreclose the other's remedies.( 44)

Equity measured at a single point in time also misses the temporal structure of disadvantage. Accessibility to jobs is worth less to a worker whose shift falls outside the transit peak, and a threshold that ignores time of day can certify access that does not exist in practice.( 13) Time-dependent measures address this by computing accessibility across the day, but they raise the data burden and are still applied in a minority of studies.( 38) The temporal gap between measured and experienced accessibility is thus a further site where normative commitment and measurement diverge.

The treatment of affordability shows how far measurement has drifted from its normative origins. Transport poverty research treats affordability as part of the distributional question, because fares and fuel costs determine whether accessibility can be used.( 35) Yet accessibility inequality indices computed from network data exclude cost, and equity studies that include fares are the exception.( 41) The omission is consequential for the Global South, where cost, not distance, is the binding constraint on participation.( 37)

The applications are broad but uneven. Hospital accessibility has been measured across Canadian and global contexts,(8,9) job accessibility across eleven Canadian regions(49) and across London, São Paulo and the Randstad,(50) and green space equity across 254 Chinese cities.(51) Latin American studies document how public transport restructuring redistributes accessibility in Bogotá(14) and how competition for opportunities shapes measured equity in Santiago.(15) Yet few of these studies state which normative principle their index implements. Table 2 summarizes the main approaches. (Table 2)

The fourth axis confronts the gap directly. The most consistent finding is negative: the philosophical and measurement literatures do not yet speak the same language. Lewis et al.(3) concluded that equity norms enter transportation research implicitly rather than explicitly, and van Wee( 4) listed the derivation of equity indicators from normative principles among the field's open questions. Ricord et al.( 52) added a data layer to the same problem, showing that biases in transportation data can silently distort equity conclusions.

A recent translation cluster begins to close the gap. Humberto et al.( 5) converted equality, priority and sufficiency into three separate metrics and applied them to Latin American cities, finding that the three principles rank the same cities differently. Lucas et al.( 53) had earlier combined ethical theories with accessibility measures to the same end, and Azmoodeh et al.( 48) built a fuzzy capability index for social inclusion. These studies share a design: name the principle first, then choose or construct the indicator. The trade-off is equally clear. Composite metrics gain normative transparency but lose the simplicity and comparability of a single index, and their weights remain open to contestation.( 5)

A defensible translation must satisfy three conditions, and the translation studies differ on all three. It must state the principle, map it to an indicator, and expose the indicator's sensitivity to its own parameters.( 5) Lucas et al.( 53) satisfied the first two conditions but left sensitivity implicit, whereas Humberto et al.( 5) demonstrated that the three principles diverge once applied to the same cities. Until these conditions are standardized, translations will remain persuasive demonstrations rather than transferable tools.

The governance of translation is the last unresolved piece. Who chooses the principle, the threshold and the weights is a political question that the technical literature leaves open.( 20) Ryan et al.( 23) note that no agency has defined a sufficient level of accessibility, which implies that the decision has no institutional owner. Until ownership is assigned, translation metrics will be adopted ad hoc by researchers and will not bind appraisal.( 21)

Policy evidence points in the same direction. Where equity is an explicit objective, measurement follows; where it is not, distributional effects go unreported. Guzmán et al.( 14) showed that Bogotá's integrated transit reforms changed accessibility in ways that were legible only once equity was measured, and Craik et al.( 54) estimated the equity impacts of the London congestion charge with a design that required an explicit counterfactual. Knap et al.( 12) linked a cycling accessibility metric to spatial and socioeconomic inequality in Utrecht, and Jomehpour Chahar Aman and Smith-Colin( 55) mapped transit deserts as a needs-based measure. The pattern is consistent: normatively explicit studies produce actionable findings, while normatively silent indices remain decorative.

The institutional obstacle is as important as the technical one. Translation studies remain demonstrations rather than routines, because appraisal guidance still rewards travel time savings and because equity indicators are added after the fact instead of shaping the option set.( 6) Where agencies have adopted a standard, as in the needs-gap approach of transit desert mapping,( 55) the indicator has moved from research into operation. The bottleneck is therefore not the absence of a metric but the absence of a mandate.( 21)