In June 2025, the MTA redesigned the bus network in Queens and eliminated the QM3 express line serving the eastern part of the borough. Certain lines in the northeast and southeast received improvements; the QM3 from Little Neck was discontinued, with local and rail alternatives offered. The logic was one of optimization: concentrate resources where demand is measurable. This is precisely where the problem begins.

The Essential Points

  • Agencies redesigning their networks based on ridership data reproduce existing geographic inequalities: underserved areas generate little signal, therefore little investment.
  • The Queens bus network redesign (MTA, June 2025) illustrates this logic: the northeast and southeast benefited from improvements, the east lost its express QM3 route (Eno Center for Transportation).
  • Chicago, Vancouver, and Toronto are conducting similar restructurations according to the same methodology based on observed demand.
  • Alternatives exist: some agencies integrate synthetic mobility data or structured community consultations to compensate for statistical blind spots.
  • Governance remains a central issue: decision-makers choose the indicators and are responsible for translating equity objectives into operational criteria.

The Logic of Ridership Data Creates Its Own Blind Spots

A bus network is not built from a blank slate. It inherits accumulated decisions made over decades: historical routes, residential density, the presence or absence of heavy infrastructure. When an agency launches a redesign, it starts from what data tells it about current travel behaviors. Boardings and alightings, passage frequencies, load factors by line: these metrics are concrete, comparable, defensible before a board of directors.

The problem is that this data measures expressed demand, not latent demand. In historically underserved areas, residents may adapt their travel patterns to the absence of reliable transportation, for example by shortening certain trips, relying on informal ride-sharing, or forgoing certain jobs or services. Low ridership may reflect insufficient service rather than an absence of need. A standard optimization algorithm does not make this distinction. It sees few passengers and concludes that the line deserves fewer resources.

The Eno Center analyzes redesign practices and trade-offs between ridership and coverage, without establishing in this source the specific dynamics described. A methodology based exclusively on past ridership can consolidate existing imbalances rather than correct them.

Queens, June 2025: Three Zones, Three Trajectories

The Queens case concretely illustrates this mechanism. The MTA presented its redesign plan as a modernization of the network, aimed at improving commercial speed and clarity of service. For the northeast and southeast of the borough, the results are tangible: increased frequencies, better-articulated connections, reduced travel times.

Eastern Queens experienced a different fate. The QM3 express line, which connected residential neighborhoods to Midtown Manhattan, was discontinued. The discontinuation may impose a connection transfer on some former users; the MTA did identify local alternatives and the LIRR, however, without publishing here a generalized comparison of travel times. The announced alternatives are the Q12, Q13, Q36, or the LIRR; no official comparison of commercial speed was found.

The discontinuation fits within a logic of allocation based on ridership volumes. The MTA motivated the discontinuation primarily through the low ridership of the QM3, within a broader restructuring and service reallocation trade-off; it did not publish a justification based on pure efficiency calculation. An equity analysis might conclude that replacing the QM3 imposes additional connections, costs, or access constraints on certain residents. Conversely, it would be inaccurate to assert without detailed analysis that there is no fast alternative to Manhattan: the MTA notably identified the LIRR.

The distinction matters because it touches on who bears the costs of optimization. Frequency gains in the northeast and southeast benefit their residents. Degradation in the east is borne by others. Both decisions result from the same redesign exercise, presented as a net benefit for the network.

Chicago, Vancouver, Toronto: A Continental Trend

Queens is not an isolated case. The Eno Center publishes analyses and contributes to research on redesigns, without a systematic continental follow-up being demonstrated by the consulted source. These agencies plan network modifications using data and consultations, but their methodologies and objectives are not identical.

The results present similarities. In each of these cities, certain corridors gain in service quality, higher frequencies, expanded schedules, better connections with heavy transit lines. In other areas, service may shrink or disappear. Sectors losing service may be located on the periphery of existing networks, where baseline ridership may be more modest.

This observation raises a methodological question that the agencies themselves are beginning to formulate. A redesign should integrate not only observed demand, but also the demand that the current network has discouraged or made impossible. This distinction determines whether a redesign improves system equity or simply redistributes efficiency within a preexisting structural imbalance.

The analogy with other digital resource allocation systems is direct. A model that learns from past behaviors reproduces the constraints that shaped those behaviors. This is an issue found in very different contexts, from AI deployment in enterprise to logistics of merchandise flows, and arises here in its most concrete form: daily access to the city.

Agencies Seeking to Correct the Ridership Bias

These three approaches share a common limitation: they intervene in a process whose frame of reference remains defined by available data. Integrating complementary sources or weighting equity criteria improves model precision but does not change the relationship between the agency and the affected populations. Alternative indicators are still selected and interpreted by the same technical teams according to their own conventions. Community consultation, when introduced late, informs without constraining. This asymmetry delimits the scope of methodological corrections: they make the model less blind without making the decision-making process more permeable to the trade-offs that communities would like to make themselves.

Some agencies have begun to work on this problem with varying degrees of sophistication. The first approach consists of integrating complementary mobility data beyond raw ridership: mobile phone data, origin-destination surveys, pedestrian flow analysis at stops. These sources allow estimation of potential mobility in low-ridership areas without limiting to current passengers.

The second approach is institutional. Several agencies have introduced explicit equity criteria in their evaluation frameworks for redesigns. This can take the form of minimum service thresholds by geographic zone, weightings applied to sectors identified as disadvantaged, or access objectives for employment and services measured independently of ridership. The Régie de transport métropolitain of Montreal, for example, publishes spatial equity analyses in its planning documents.

The third approach is participatory. Structured community consultations, conducted before finalizing redesign plans, can surface needs that aggregate data does not capture: trips to health facilities, access to specific markets, mobility of workers with atypical schedules. This method is costly in time and resources, and its results are not always easy to integrate into a modeling tool. But it fills a blind spot that ridership data alone cannot correct.

These approaches are not mutually exclusive. Their effectiveness depends mainly on their integration early in the redesign process, not as a correction added after major decisions have been made.

The Indicators Absent from the Calculation Determine What the Network Does Not Finance

Behind technical choices about metrics lie political choices about priorities. An agency that defines its network’s success by cost per passenger-kilometer will optimize toward dense lines. An agency that adds an indicator for employment access for carless households will make different decisions, sometimes more costly in appearance, but which otherwise distribute the benefits and burdens of the system.

This tension between efficiency and equity is not unique to transportation. It runs through all urban public services that rely on data to allocate resources. Analysis of global supply chains shows how systems designed for optimization can create invisible fragilities until they materialize. In urban transportation, the fragility is social rather than geopolitical: it manifests in longer travel times, inaccessible jobs, neighborhoods that break out of isolation more slowly than their neighbors.

The Queens network redesign illustrates a concrete governance problem: local elected officials are rarely involved in defining modeling metrics. User associations may comment on plans, but rarely upstream of modeling. Affected communities often discover decisions at the stage of formal consultation, when major orientations are already set.

Agencies like Transport for London have developed more robust equity evaluation frameworks, with regular audits of territorial access to services. These frameworks do not eliminate difficult trade-offs but make them explicit and contestable. The difference is significant: an opaque trade-off maintains the status quo; a visible trade-off can be discussed, amended, reviewed.

Making Trade-offs Visible to Make Them Contestable

Visibility of trade-offs also supposes that agencies clearly distinguish what constitutes a real budget constraint from what constitutes a prioritization choice. The two produce service cuts but do not have the same political status. A constraint endured reduces the field of possible decisions; a prioritization choice opens it to deliberation. As long as public documents present the two as equivalent, affected populations do not have the elements necessary to evaluate whether the decision was inevitable or contestable. Making this gap legible is a prerequisite for any governance that claims to integrate equity as an operational criterion rather than a declarative objective.

The Queens case does not argue against network redesigns. Restructurations can improve service on certain axes, reduce operating costs, and make networks more legible. Some case studies report gains on reinforced corridors; no continental synthesis found allows generalizing this result.

The issue is to make visible the zones bearing the costs of optimization and to create mechanisms so these costs are deliberately weighed rather than simply absorbed. This supposes that agencies publish not only their ridership projections but also their impact analyses by geographic zone and income group before decisions are made.

Several questions remain open after the Queens experience. The MTA plans consultation and a public hearing for major changes; the source does not describe a specific post-discontinuation appeal mechanism for the QM3. Agencies do not systematically publish the evolution of service in zones that lost lines two or three years after a redesign. The MTA plans ridership monitoring and frequency adjustments on certain lines; no specific mechanism for restoring the QM3 based on latent demand was identified.

Answers to these questions do not change the fundamental logic of redesigns. They determine whether these redesigns serve the entire city or only the part of the city that data already knows how to count.

Sources

  1. Eno Center for Transportation, Bus Network Redesigns in the Modern Age: https://enotrans.org/article/bus-network-redesigns-in-the-modern-age-how-u-s-transit-agencies-adapt-to-evolving-travel/
  2. MTA, Queens Bus Network Redesign Implementation Plan, June 2025 (Metropolitan Transportation Authority, New York)
  3. Transport for London, Equality Impact Assessment Framework (Transport for London, annual planning report)