How Town Planning Works | TPW-0110 — The Concurrency Test: How Adequate Public Facilities Decide Whether Growth Can Proceed Before Infrastructure Catches Up

Series ID: TPW-0110

A city can approve four thousand new homes on paper while the sewer system can physically serve only fifteen hundred of them.

It can also make the opposite mistake.

One road segment reaches an old congestion threshold, the planning system declares capacity exhausted, and a transit-oriented housing project is delayed even though most future residents could walk, cycle or use the station beside it.

Both failures come from the same planning problem: land-use approval and infrastructure capacity are being treated as separate worlds.

Adequate public facilities systems—often called APF, adequate public facilities ordinances, or concurrency systems in some jurisdictions—try to connect those worlds. The basic idea is simple: growth should proceed when the facilities needed to serve it are available, funded, programmed or otherwise capable of meeting an adopted service standard.

The American Planning Association’s April 2026 Modernizing Adequate Public Facilities Practices issue describes the principle directly: development approvals are linked to transportation, water, sewer, schools, parks, emergency services or other facilities so that increased demand does not push service below an acceptable level. APA also highlights the modern problem. Older APF systems can become barriers to needed housing when they rely on rigid pass-fail tests, outdated data or single-mode measures that no longer match current goals such as complete streets, transit, attainable housing, climate resilience and environmental justice.

That tension is the subject of this article.

The reader job: connect growth approval to infrastructure without turning capacity into a permanent veto

This article explains how concurrency works as a planning mechanism: what capacity means, how levels of service are set, how capital plans connect to development approvals, how capacity is reserved, how cumulative demand is counted, how fees differ from timing tests, and how a modern system can support housing and multimodal growth instead of merely saying “full”.

Neighbouring mechanisms already have owners. Density and Capacity owns the broader question of how much town can fit before systems break. The Sewer Capacity Map owns wastewater constraints in detail. The Time Layer owns development sequencing. The Financial Machine Behind the Map owns the relationship between land, infrastructure and finance.

The Concurrency Test owns a narrower question: when a development adds demand to shared public systems, how should the planning authority decide whether that demand can be carried now, should wait for a funded improvement, or should trigger a proportionate response?

Concurrency is a timing rule disguised as a capacity rule

At first glance, concurrency sounds like a capacity question: is there enough road, sewer, water, school or park capacity?

The more useful interpretation is temporal.

The facility may be adequate today but inadequate after two thousand approved homes are occupied. It may be inadequate today but fully adequate after an upgrade already funded for next year. A wastewater plant may have treatment capacity while the trunk pipe serving one district is full. A school system may have spare seats regionally but none near the proposed neighbourhood.

So the real question is not simply “Is capacity available?” It is: when will capacity be available relative to when this development creates demand?

That makes concurrency part of the town’s sequencing machinery.

The first decision is which facilities belong in the test

Not every public service needs to be governed through concurrency.

Common facility categories include potable water, sanitary sewer, stormwater or drainage, transportation, schools, parks, fire and emergency response, solid waste and sometimes other locally important infrastructure.

The legal list varies. Florida provides a useful current example because its statutory framework distinguishes mandatory statewide concurrency subjects from optional local ones. The Florida Commerce Capital Improvements Element guidance states that sanitary sewer, solid waste, drainage and potable water remain facilities subject to statewide concurrency requirements, while parks and recreation, public schools and transportation may be handled through optional local concurrency systems.

That Florida structure should not be copied blindly elsewhere. Its value as an example is institutional: the jurisdiction must decide which facilities require a formal timing test and which should be managed through other planning, funding or service-delivery tools.

Every facility needs its own meaningful measure of service

Capacity is not one universal number.

A water system may be limited by source supply, treatment capacity, storage, pressure, pipe diameter or drought reliability.
A sewer system may be limited by pipe capacity, pump stations, treatment capacity, wet-weather infiltration or discharge permits.
A school system may track seats, class sizes, catchments, specialised programmes and travel distance.
A fire service may care about response time, station coverage, staffing, apparatus and road access.
A park system may examine acreage, walking access, facility type, quality and population served.
A transport system may measure vehicle delay, person throughput, travel time, safety, transit reliability, pedestrian conditions or several modes together.

A concurrency system becomes only as intelligent as the metric it chooses.

A bad metric can make good planning fail

APA’s 2026 APF guidance highlights a central weakness of older systems: many were built around single-mode measures and binary pass-fail thresholds.

Transportation provides the clearest example.

If the only measure is vehicle delay at an intersection, a dense mixed-use station area can perform badly because cars move slowly—even while large numbers of people walk, cycle or use transit efficiently. A low-density edge project may perform better on the same metric because traffic is dispersed, even though it generates longer car trips and larger infrastructure demands over time.

The measurement system can therefore reward the development pattern the comprehensive plan is trying to avoid.

A modern concurrency framework should ask whether its metric measures the public outcome the community actually values.

Level of service is a policy choice before it becomes a technical calculation

Most APF systems rely on an adopted level of service, often abbreviated LOS.

The phrase sounds objective. The number is not discovered in nature.

A city chooses what service standard it considers acceptable. That choice reflects cost, risk tolerance, equity, desired urban form and public priorities.

A road can always be made to carry fewer cars more quickly if enough land and money are devoted to widening it. A school system can always target smaller classes if taxpayers fund the additional buildings and staff. A park system can target more acreage if land is available. A water utility can build additional redundancy at a cost.

The adopted standard is therefore a political and planning commitment about what the community is prepared to provide and maintain—not a theoretical maximum.

The capacity test should use adopted policy, not abstract perfection

If every development must wait until every public system is ideal, almost no city could grow.

The test should compare projected demand against the adopted service standard. If the current system already operates below that standard, the plan should identify the deficiency and the capital strategy to correct it.

This is why concurrency cannot be separated from capital planning.

The capital improvement plan is the other half of concurrency

A capacity rule without a capital plan becomes a brake.

If the sewer is full and the city has no funded plan to expand it, development waits indefinitely. If the sewer is full but a funded trunk upgrade will open in eighteen months, the city can phase approvals around that known event.

Florida’s current Capital Improvements Element guidance illustrates this connection. The five-year capital improvements schedule identifies projects needed to implement the comprehensive plan and maintain adopted service levels. The concurrency mechanism and the capital budget are therefore meant to tell the same story.

This is the deeper planning rule: do not promise land-use capacity the capital programme cannot support, and do not build capital infrastructure with no land-use strategy to use it.

Existing capacity, committed capacity and speculative capacity are different

Capacity accounting needs categories.

Existing available capacity: service that physically exists and is not already committed.
Committed capacity: capacity reserved for approved or contracted development that has not yet generated full demand.
Funded future capacity: improvement included in an adopted, financially credible programme.
Programmed but uncertain capacity: a project in a plan without secure funding, land, approvals or delivery schedule.
Speculative capacity: an improvement that somebody hopes will happen.

These categories should not be treated as equivalent.

The more uncertain the facility, the more carefully the planning authority should decide whether development may rely on it.

Capacity reservations prevent the same seat, litre or road space from being sold twice

Imagine a sewer basin with capacity for another 1,000 homes.

Project A is approved for 700. Project B arrives six months later for 600. If the city looks only at current flows, the pipe may still appear to have room because Project A is not built yet.

Approving Project B without accounting for A means the authority has allocated the same future capacity twice.

A reservation ledger solves that problem. Once a project reaches the point defined by local law, some or all of its projected demand is treated as committed. The system then shows physical capacity minus existing demand minus valid reservations.

Approved-but-unbuilt development belongs in the ledger

Large planning pipelines can hide enormous future demand.

A city may have 20,000 approved homes but only 8,000 occupied. Utility flows reflect the occupied population. Infrastructure planning must also understand the entitled pipeline.

The challenge is that not every approved project will be built. Reserving capacity forever for dormant approvals can create artificial scarcity.

The solution is an explicit reservation policy: when capacity is reserved, how long the reservation lasts, what milestones keep it alive, and when unused capacity returns to the pool.

Cumulative impact is why the last project through the door should not pay for the whole problem

Infrastructure constraints accumulate.

Ten projects may each add modest demand to a road, school or wastewater plant. The tenth application may be the one that crosses the adopted threshold, but it did not create the first nine shares of demand.

A fair system distinguishes marginal impact from cumulative deficiency.

This is one reason proportionate-share mechanisms exist. The final applicant should not automatically be asked to build an entire regional facility merely because it arrived when the remaining capacity reached zero.

Concurrency and impact fees answer different questions

An impact fee asks: how should new development contribute financially toward facilities needed because of growth?

A concurrency test asks: will the facility be adequate when the development needs it?

These questions interact but they are not interchangeable.

A developer can pay a fee today while the road, school or pipe is still ten years away. Money in an account does not create physical capacity on the required date.

Conversely, a public facility may already have adequate capacity even though the jurisdiction still lawfully charges a growth-related fee for future system expansion.

A fee does not create a pipe instantly

This sounds obvious but is one of the most important concurrency principles.

Infrastructure requires land, design, permits, procurement, construction and commissioning. A treatment plant may take years. A school needs a site. A road project may depend on another agency. A new fire station requires staffing after the building opens.

The concurrency system therefore needs a schedule, not only a funding mechanism.

Proportionate share can convert failure into a route forward

Some systems allow an applicant whose project would contribute to a facility deficiency to fund or construct a proportionate share of the needed improvement.

Florida’s current transportation concurrency guidance requires local systems that retain transportation concurrency to allow proportionate-share contributions consistent with state law.

The planning value of proportionate share is not that every project can buy its way through a failed test. The value is that the system can connect development demand to a defined improvement instead of producing only a binary denial.

The improvement still has to be real, deliverable and sufficient.

The infrastructure project must be deliverable

A capital plan can contain a project that exists mainly as a line in a spreadsheet.

Before development relies on it, ask whether the facility has funding, land, design progress, permits, institutional responsibility and a plausible construction schedule.

A bridge with no right of way, a school with no site or a treatment plant with no financing should not be treated as equivalent to a project under construction.

Concurrency becomes trustworthy when future capacity has a confidence level.

Housing goals should change how the system is designed

APA’s 2026 APF guidance explicitly warns that outdated systems can become significant barriers to needed housing.

That does not mean housing should ignore infrastructure constraints.

It means the concurrency framework should distinguish a genuine physical bottleneck from an obsolete metric, a funding delay, an overly conservative assumption or a facility standard that conflicts with adopted housing policy.

If the comprehensive plan calls for 50,000 additional homes but the capital programme funds infrastructure for 20,000, the contradiction is not the fault of the 20,001st housing application. The plan and budget are misaligned.

Affordable housing is especially sensitive to delay

Delay has carrying cost.

Land loans, consultant teams, option agreements, tax-credit deadlines, grants and construction prices continue moving while a project waits for capacity.

A concurrency system that places needed affordable housing in an indefinite queue can undermine the public goal it was supposed to serve.

Possible policy responses include priority capacity, discounts or exemptions where legally permitted, public funding of strategic infrastructure, or separate service standards for designated growth areas. The correct tool depends on local law and the nature of the constraint.

Infill and greenfield growth create different infrastructure economics

An infill project may use existing roads, pipes, parks and schools, but those systems can be old or locally constrained.

A greenfield project may have no local infrastructure at all, yet new systems can sometimes be designed comprehensively from the start.

A simplistic capacity test can therefore produce surprising results. The edge site passes because a new road is easy to size, while the infill site fails because one existing junction is congested.

The planning system should ask which development pattern creates lower long-term public cost and better access—not only which project produces the easiest short-term engineering calculation.

Schools should be treated as a network, not a single seat count

A district may have spare school capacity overall while the neighbourhood school beside the project is full.

Another area may have empty seats but require long travel or cross unsafe roads. Specialised programmes, age groups and school-choice policies can complicate the picture further.

A school concurrency system should therefore define the geographic scale at which capacity is measured and how boundary changes, expansions, new schools and transport are considered.

The metric should reflect educational access, not merely arithmetic vacancy.

Parks need more than acreage

Ten hectares of inaccessible wetland and ten hectares of neighbourhood playing fields are not equivalent recreation capacity.

Park planning can consider walking access, type of facility, quality, usable area, programming, ecological role and population served.

A purely per-capita acreage standard can encourage land acquisition without asking whether residents can actually use the land.

Modern APF systems should connect quantity to function.

Fire and emergency capacity often depends on response time

A fire station does not serve a fixed number of houses in the same way a treatment plant serves a flow volume.

Road geometry, travel time, staffing, apparatus, simultaneous incidents and barriers all matter.

A development at the edge of the service area may create a coverage problem even when systemwide staffing appears adequate. Conversely, a compact infill project near an existing station may add population without materially worsening response performance.

The facility metric should match how the service actually works.

Water capacity contains several bottlenecks

A city can have enough water rights but not enough treatment capacity.

It can have treatment capacity but inadequate storage. It can have systemwide storage but low pressure in one elevated district. It can have ordinary-year supply but insufficient drought reliability.

Florida’s current Water Supply Planning guidance shows how capital planning, water-supply plans and adopted service standards must be coordinated over time.

A concurrency test should therefore identify the binding constraint rather than treating “water capacity” as one number.

Wastewater has dry-weather and wet-weather capacity

A sewer system may operate normally on a dry day and overflow during heavy rain because groundwater and stormwater enter old pipes.

That means average daily flow is not always the correct capacity measure.

The concurrency test may need to account for peak flow, pump-station limitations, wet-weather infiltration and treatment constraints. The existing Sewer Capacity Map explains this technical chain in detail.

Transportation measures should be multimodal where the plan is multimodal

Florida’s current transportation guidance explicitly allows local alternatives to traditional roadway concurrency, including area-wide service measures, multimodal standards, complete-street strategies and systems that give greater priority to pedestrian environments and transit.

This is an important modern signal.

If a community wants compact mixed-use growth near transit, its infrastructure test should not automatically penalise that growth because cars experience urban conditions.

Metrics can include person throughput, travel time, transit frequency, pedestrian connectivity, bicycle comfort, safety and network accessibility rather than relying only on vehicle speed.

A failed capacity test should trigger options, not only denial

When a project fails concurrency, the planning system should be able to explain the route forward.

Delay occupancy until a funded improvement opens.
Phase the project so early demand fits existing capacity.
Reduce or redesign the development.
Construct or fund a proportionate improvement.
Shift demand through transport, water-efficiency or operational measures.
Reserve capacity under a defined milestone system.
Accelerate a public capital project where the development advances adopted priorities.
Use a different site or service arrangement where the bottleneck cannot reasonably be solved.

A binary no-answer is sometimes necessary. It should not be the only vocabulary the system knows.

Concurrency should not become a disguised no-growth policy

A jurisdiction can set service standards so high and fund improvements so slowly that almost all development fails.

That may look technical while functioning as a growth moratorium.

If the community genuinely wants to stop growth, that policy should be debated transparently through the plan and law. Capacity standards should not be manipulated to achieve a hidden policy objective.

The test is consistency: do the housing, economic-development and growth policies in the plan align with the service standards and capital budget that determine whether those policies can be built?

Publish capacity before applications arrive

Applicants should not discover infrastructure failure after spending years on design.

A public capacity map or dashboard can identify constrained sewer basins, school areas, water zones, road networks or fire-service gaps. It can show existing demand, committed demand, funded improvements and expected dates.

The information will not be perfect. Publishing it early lets land markets, public agencies and developers make better decisions before money is sunk into impossible assumptions.

Capacity data must have an age

A number without a date is dangerous.

Population changes, approvals accumulate, projects expire, utility flows vary and capital works open. School enrolment shifts. Transit service changes. A fire station closes for rebuilding.

Every published capacity value should carry an observation date, methodology and update frequency.

Old capacity data can be worse than no data because it creates false confidence.

Forecast uncertainty should be explicit

Future demand depends on household size, water use, school-age population, travel behaviour, economic activity and many other variables.

One development of 1,000 homes may not create the same school, water or traffic demand as another 1,000 homes.

Use scenarios rather than one apparently precise forecast where uncertainty is material. A range can show how much capacity exists under low, central and high demand assumptions.

The goal is not to eliminate uncertainty. It is to stop hiding it inside a single number.

Capacity can be expanded operationally, not only with concrete

Sometimes the cheapest new capacity comes from operating the existing system better.

Water conservation can reduce demand. Infiltration repairs can recover sewer capacity. Bus priority can increase person throughput. School boundary adjustments can rebalance seats. Signal coordination can improve network performance. Staffing changes can improve emergency response.

The concurrency system should not assume every deficiency requires a new megaproject.

Maintenance belongs in the capacity model

A bridge, pump station or water main that exists on the asset register may not be reliable capacity if it is near failure.

Deferred maintenance can therefore create hidden concurrency risk.

The existing Maintenance Ledger explains why cities need to track inspection, repair, renewal and replacement. A capacity system should use that asset condition information rather than counting every installed facility as equally dependable.

Climate resilience changes the definition of adequate

A stormwater system sized to yesterday’s rainfall may appear adequate on an old standard and fail under current conditions.

A water system with enough average supply may be vulnerable during longer droughts. A coastal road may technically provide access but become unreliable during repeated flooding. A fire-service model may change as wildfire exposure expands.

Modern concurrency should therefore consider whether adopted service standards remain credible under changing hazards.

Worked example: 4,000 homes beside a station

Imagine a city plans a new mixed-use station district with 4,000 homes.

The transport plan supports the location. Water capacity is adequate. The school authority can absorb the first phase. The wastewater basin, however, has physical capacity for only 1,500 additional homes until a trunk sewer upgrade opens.

A crude concurrency system says the 4,000-home plan fails.

A better system separates entitlement from occupancy timing. It can approve the land-use framework, reserve existing sewer capacity for an initial phase, connect later permits to the funded trunk upgrade, allow the developer to contribute proportionately or accelerate the project, and update reservations as phases are built.

The infrastructure constraint remains real. The planning answer becomes sequencing rather than abandonment.

The capacity queue needs priority rules

Scarce capacity creates a queue.

Who gets the last 500 sewer connections? First application submitted? First building permit issued? Affordable housing? A hospital? A project inside the designated growth area? A development that funds the upgrade?

If the answer is not defined, capacity allocation can become opaque and political.

Priority policy should be adopted before scarcity becomes acute. The rule can reflect public objectives, legal rights, application milestones and infrastructure strategy.

A temporary bottleneck is different from structural scarcity

Suppose a treatment plant upgrade is under construction and will open in nine months.

That is a temporary bottleneck. Phasing or delayed occupancy can solve it.

Now suppose the water source is fully allocated and no feasible new supply exists for decades.

That is structural scarcity. The land-use plan itself may need to change.

Concurrency should distinguish the two because the appropriate planning response is completely different.

Interim solutions need an end date and performance standard

Temporary facilities can bridge a gap: modular classrooms, interim water storage, shuttle buses, temporary fire facilities or onsite treatment.

Temporary solutions can also become permanent through inertia.

If an interim measure is used to satisfy concurrency, specify its capacity, safety standard, monitoring, responsible operator and end condition. State what permanent facility replaces it and when.

The service provider and planning authority need a shared ledger

The planning department may approve development. The utility authority controls sewer. The school board tracks enrolment. The transport agency runs capital projects. The fire service owns response data.

Concurrency fails when these systems use different project lists, dates and assumptions.

A shared ledger should connect development applications, approved entitlements, reservations, actual occupancy, current facility capacity and funded improvements.

The infrastructure state should not depend on one planner emailing five agencies for spreadsheets every time an application arrives.

Data governance matters because capacity numbers move money

A capacity figure can determine whether land can be developed, whether a project is delayed and whether millions of dollars of infrastructure are required.

That makes the data economically consequential.

Define who owns each dataset, who may change it, how corrections are logged, how often it is refreshed and what evidence supports an applicant challenge.

Capacity should be auditable. A developer should not be told “the system is full” without a traceable calculation.

Scenario planning prevents the city from building the wrong infrastructure

Capital plans often assume one growth forecast.

But housing production, employment, technology and travel behaviour can diverge from forecasts. Building a large road expansion around one high-car scenario can lock in a development pattern the plan does not want.

Test several growth patterns and service strategies. Which investments are robust across scenarios? Which depend heavily on one forecast? Which can be phased as demand becomes clearer?

Concurrency becomes more resilient when the capital programme is adaptable rather than overconfident.

A designated growth area should have a capacity strategy before the city markets it

Cities sometimes designate a corridor or station area for major housing growth and then discover that sewer, schools or power cannot support it.

That reverses the correct sequence.

Before inviting private investment, identify the infrastructure needed for the planned build-out, likely funding sources, delivery stages and critical bottlenecks.

A growth map without a capacity strategy is a promise without an operating plan.

Fairness is larger than the question of who pays

Concurrency distributes more than infrastructure cost.

It distributes delay, development opportunity and access to scarce capacity.

If wealthy projects can finance upgrades while smaller builders cannot, capacity rules may concentrate development opportunity. If infill areas inherit old infrastructure deficiencies while greenfield sites receive new systems, the framework may push growth outward. If existing residents receive upgraded facilities paid entirely by new housing, distributional questions arise in the other direction.

Equity analysis should therefore examine who benefits, who waits and who pays—not merely whether the accounting balances.

A concurrency failure is also a budget signal

When strategically important projects repeatedly fail for the same facility, the problem may belong in the capital budget rather than the development application.

Repeated sewer failures in the designated housing corridor say: fund sewer capacity. Repeated pedestrian-access failures around transit say: build the missing walking network. Repeated school shortages in the growth area say: the education capital programme needs land and funding.

The permit system is producing information about public investment priorities.

Worked example: the wrong metric blocks the transit project and rewards the edge

Imagine two proposals of 500 homes.

Project A sits beside a frequent rail station in an existing mixed-use district. Streets are congested at peak hour, so an intersection-delay test fails.

Project B sits on the urban edge beside a wide road with spare vehicle capacity. It passes the same test.

Yet Project B creates longer trips, greater car dependence, new utility extensions and larger future maintenance obligations. Project A places more residents near transit, jobs and existing services.

If the plan says compact transit-oriented growth is the goal, the concurrency metric is contradicting the plan.

The solution is not to pretend Project A has no transport impact. It is to measure mobility in a way that recognises person movement, mode choice, safety and the network outcomes the adopted plan actually values.

A concurrency audit

Facilities: Which public systems are formally subject to the concurrency test?
Authority: What law or adopted plan authorises the test?
Service standard: What level of service is adopted for each facility?
Metric: Does each metric represent the public outcome that actually matters?
Geography: At what spatial scale is capacity measured?
Existing demand: What portion of capacity is already used?
Committed demand: How are approved-but-unbuilt projects counted?
Reservations: When is capacity reserved and when can a reservation expire?
Future capacity: Which improvements are funded, designed, permitted or merely proposed?
Timing: Will the facility be available when the development generates demand?
Cumulative impact: How is demand from several developments combined?
Proportionate share: Can a project contribute fairly to a defined improvement?
Impact fees: Are funding obligations distinguished from timing and physical capacity?
Pipeline: Does the model account for dormant approvals without reserving capacity forever?
Housing: Does the system support or frustrate adopted housing-production goals?
Infill: Are existing urban areas penalised by metrics designed for edge growth?
Modes: Does transport concurrency recognise transit, walking and cycling where policy requires them?
Data age: When was the capacity information last updated?
Uncertainty: Are forecasts shown as ranges where appropriate?
Maintenance: Does asset condition reduce dependable capacity?
Climate: Do future hazards change what adequate service means?
Queue: How is scarce capacity allocated fairly?
Remedies: What options exist after a failed test besides denial?
Capital feedback: Are repeated failures translated into public investment priorities?
Transparency: Can applicants and residents reproduce the capacity calculation?

The land-use plan and the capital budget must tell the same story

A city cannot plan housing for one place, jobs for another and growth for a third while its pipes, schools, roads and emergency services are funded according to an unrelated map.

Concurrency exposes that contradiction because every development application eventually asks the same practical question: can the systems carry what the plan says should happen here?

A good APF system does not make growth wait forever for perfect infrastructure.

It identifies the real bottleneck. It measures the right outcome. It counts existing and committed demand honestly. It distinguishes funded capacity from wishful capacity. It connects thresholds to capital projects. It provides routes through phasing, operational improvement and proportionate mitigation. It protects scarce capacity from double counting. It updates its data. It learns when the same failure repeats.

Most importantly, it forces the city to align two powerful machines that are too often separated: the machine that grants development rights and the machine that builds and maintains public infrastructure.

The purpose of concurrency is not to stop the town from growing. It is to make growth and infrastructure arrive in an order the town can actually carry.