How Town Planning Works | The Density Measure

Series ID: TPW-0158

“This neighbourhood is too dense.”

“That development is not dense enough.”

“We need thirty dwellings per hectare.”

“The site already has a plot ratio of 2.8.”

All four statements can sound as if they are talking about the same thing.

They are not.

Density is one of the most frequently used words in town planning and one of the easiest to misuse. Sometimes the numerator is dwellings. Sometimes it is people. Sometimes it is bedrooms, jobs or floor area. Sometimes the denominator is the whole master-planned area, including streets and parks. Sometimes it is only the residential development parcels. Sometimes the number is quoted per hectare. Sometimes per acre. Sometimes a planning document uses a minimum density to prevent underdevelopment; another uses a maximum to control intensity.

Then the argument becomes even more confused because people use “density” as a proxy for height, crowding, congestion, neighbourhood character, housing quality or pressure on services.

A useful density measure should do something much more modest and much more precise: state exactly what is being counted, exactly what land area it is being divided by, and exactly what planning decision that ratio is meant to inform.

Current planning systems continue to rely on this distinction. London’s planning framework treats density through a design-led approach to optimising site capacity rather than a single universal number. Local plans in England still use dwellings per hectare in location-specific ways. Planning guidance in several jurisdictions distinguishes gross from net residential density because the land included in the denominator can change the apparent number dramatically. Around the world, residential density is still commonly expressed as dwelling units per hectare or acre, while population density uses people per unit of land.

The arithmetic is simple.

The planning is not.

The reader job: calculate the number correctly before using it to make a planning argument

This article owns the measurement problem: gross density, net density, dwelling density, population density, site density, neighbourhood density, unit conversions, mixed-use denominators and the difference between density and other development controls.

It does not own total floor-area capacity. That belongs to The Floor Area Ratio. It does not own minimum parcel size, which belongs to The Minimum Lot Size. It does not own ground coverage, which belongs to The Lot Coverage Ratio. It does not own the horizontal building envelope, which belongs to The Setback Line, or the vertical envelope, which belongs to The Building Height Rule.

The Density Measure answers a narrower but essential question: what does the number actually mean?

Every density number has a numerator and a denominator

At its simplest:

Density = quantity ÷ land area

The formula becomes useful only after both terms are defined.

If the quantity is dwellings, the result is dwelling density.

If the quantity is people, the result is population density.

If the quantity is jobs, it is employment density.

If the quantity is floor area rather than a count of people or units, the measure starts moving toward floor-area ratio or plot ratio, which is a different planning control.

The denominator is equally important. Ten hectares of residential parcels is not the same denominator as ten hectares containing those parcels plus a school, park, arterial road, drainage reserve and community facility.

Dwellings per hectare is not the same thing as people per hectare

Suppose two developments each contain 100 homes on two hectares.

Both have a dwelling density of fifty dwellings per hectare if the same denominator is used.

Development A averages 1.6 residents per home. Its estimated residential population is 160, or eighty people per hectare.

Development B averages 3.2 residents per home. Its estimated residential population is 320, or 160 people per hectare.

The dwelling density is identical. The population density is double.

This matters because many service demands are generated by people rather than doors. Water consumption, waste generation, school-age population, transit use and park demand are influenced by household composition, age and behaviour, not merely the number of dwelling units.

Density measurement should therefore distinguish the stable physical count of homes from the changing social reality of who occupies them.

Household size changes over time while the building stays still

A neighbourhood can lose population without demolishing a single home.

Children grow up and move out. Household sizes shrink. An older population may leave more bedrooms per person. Some units remain vacant. Other neighbourhoods move in the opposite direction as families share housing or larger households replace smaller ones.

That is why population density should not be inferred permanently from the number of dwellings.

For forecasting, planners can use an occupancy assumption:

Estimated population = dwelling count × assumed average household size

But the assumption should be stated, dated and tested against local evidence. It is a forecast parameter, not a physical characteristic of the site.

One hectare is 10,000 square metres

The hectare is the standard unit for many planning-density calculations outside the United States.

1 hectare = 10,000 square metres

A square 100 metres by 100 metres has an area of one hectare.

This visual reference helps. A density of 60 dwellings per hectare means sixty homes for each 10,000 square metres included in the chosen denominator. It does not mean every hectare must contain exactly sixty homes; a larger master plan can average that density across different parcels.

Acres and hectares need conversion before comparison

One acre is smaller than one hectare.

Approximately:

1 hectare ≈ 2.471 acres
1 acre ≈ 0.4047 hectare
1 dwelling per acre ≈ 2.471 dwellings per hectare
1 dwelling per hectare ≈ 0.4047 dwelling per acre

So a policy requiring thirty dwellings per hectare is roughly equivalent to 12.1 dwellings per acre.

A policy allowing thirty dwellings per acre is roughly 74.1 dwellings per hectare.

Confusing the units can produce a planning error of almost two and a half times.

Gross density and net density can describe the same place with different numbers

This is the central density problem.

Gross density uses a broader land area.

Net density uses a narrower development area.

The exact inclusions vary between planning systems, which is why “gross” and “net” should never be assumed to have one universal legal definition.

A gross residential calculation might include local roads, parks and community land within a neighbourhood. Another jurisdiction may include only certain supporting uses. A net calculation often focuses on the land directly assigned to residential lots or development parcels and may exclude major streets, schools, parks and other non-residential land.

The arithmetic consequence is predictable:

If the dwelling count is unchanged and the denominator becomes smaller, the density number becomes larger.

A worked gross-versus-net example

Imagine a ten-hectare new neighbourhood containing 360 dwellings.

The ten hectares include:

six hectares of residential development parcels;
1.5 hectares of local roads and paths;
one hectare of park and drainage landscape;
one hectare for a school site; and
0.5 hectare for community and utility uses.

Using all ten hectares:

Gross density = 360 ÷ 10 = 36 dwellings per hectare

Using only the six residential hectares:

Net residential density = 360 ÷ 6 = 60 dwellings per hectare

Nothing about the neighbourhood changed.

The definition of land area changed.

A density number without its denominator definition is incomplete evidence

“This district has a density of 60 dwellings per hectare” sounds precise.

It is not precise until the listener knows whether sixty is gross, net, site-level, block-level or another locally defined measure.

This is especially important when comparing cities or countries. Two planning departments can publish apparently different density standards while measuring different land components.

Comparison should therefore begin by normalising definitions, not by ranking the headline numbers.

Site density is not neighbourhood density

A single apartment parcel can have a high site density because almost all its land belongs directly to the development.

The surrounding neighbourhood can have a much lower gross density once parks, streets, schools, shops, drainage, utilities and public facilities are included.

This is not a contradiction.

The parcel is one component inside a larger system.

Town planning needs both scales. Site density helps test what a development parcel is doing. Neighbourhood density helps estimate how people and housing are distributed across the full urban fabric.

Block density adds another useful scale

A city block can be a useful denominator because it captures several parcels plus the immediate internal pattern while avoiding the much larger geography of a whole district.

But even “block” needs definition. Does the denominator stop at property lines, street centrelines or outer kerbs? Does it include rear lanes?

Density analysis becomes much more trustworthy when the map boundary is shown alongside the number.

Citywide average density can hide almost everything that matters locally

A city can contain dense centres, low-rise neighbourhoods, industrial districts, airports, forests, reservoirs and undeveloped land.

Dividing the entire population by the entire municipal area produces a valid statistical number.

It may be a weak measure of what people experience on any particular street.

This is a version of the modifiable areal unit problem: the measured value changes when the geographic unit changes.

Planning arguments should therefore use the smallest geography that matches the decision while remaining large enough to capture the relevant system.

Density is not floor-area ratio

This distinction is fundamental.

Dwelling density counts homes.

FAR or plot ratio compares floor area with site area.

Two buildings can have the same FAR and different dwelling densities because their units are different sizes.

Imagine 10,000 square metres of residential floor area on a one-hectare site.

If the average dwelling is 50 square metres before common-area assumptions, the floor area could theoretically support around 200 such unit-equivalents.

If the average dwelling is 100 square metres, it supports around 100.

The FAR is unchanged. The dwelling density changes substantially.

This is why a density policy and a floor-area policy can coexist without duplication.

Density is not building height

A tower is tall.

That does not automatically make the site denser than every low-rise alternative.

A ten-storey building on ten per cent of a large site can contain less floor area and fewer dwellings than a four-storey courtyard building covering half the site.

The Building Height Rule controls vertical envelope. Density counts a chosen activity per land area. They interact, but neither substitutes for the other.

Density is not lot coverage

A development can cover a large percentage of the ground with a low building and still contain relatively few homes.

Another can cover a small footprint with a tower and contain many homes.

Coverage tells us how much ground is occupied by building footprint. Density tells us how many selected units or people are associated with the chosen land area.

Both matter because the same density can produce very different amounts of open ground.

Density is not crowding

This may be the most socially important distinction.

A dense neighbourhood can contain spacious, uncrowded homes.

A low-density district can contain overcrowded households inside large but scarce homes.

Crowding is normally about the relationship between people and interior residential space—people per room, people per bedroom, floor area per person or another occupancy measure.

Urban density is about people or dwellings distributed across land.

Conflating them can turn a housing-quality problem into an argument against urban form.

Density is not housing quality

Quality depends on matters such as daylight, ventilation, acoustic performance, privacy, room dimensions, storage, accessibility, thermal comfort, safety, common space, maintenance and neighbourhood conditions.

A density figure does not reveal those things.

A badly designed development can be low density. An excellent one can be high density.

Planning should test quality directly rather than assuming density provides the answer.

The same density can produce radically different housing forms

Suppose the target is sixty dwellings per net residential hectare.

One hectare can be arranged as:

sixty compact rowhouses;
several low-rise apartment courts;
three medium-rise blocks;
a taller building occupying a smaller part of the site;
or a hybrid of townhouses and apartments.

The dwelling count is the same.

The street pattern, building height, open space, parking, privacy, construction method and visual character can be completely different.

This is why The Missing Middle matters. Density does not dictate one architectural type.

A minimum density and a maximum density serve opposite regulatory purposes

A maximum density prevents development from exceeding a stated intensity.

A minimum density prevents scarce urban land from being used far below the level the plan expects.

This distinction matters around infrastructure, centres and other strategically planned locations. A city can invest heavily in capacity, zone a site for substantial development, then receive a proposal for very low intensity that locks in underuse for decades.

A minimum-density policy can say that the project must reach at least a floor of residential intensity unless a stated constraint justifies less.

The policy should still be design-led. Chasing the number without regard to geometry, heritage, environmental constraints or housing type can produce poor outcomes.

A density range is often more useful than one mandatory point

One exact density can create false precision.

A range allows site conditions and design quality to influence the final result while preserving the intended order of magnitude.

For example, a policy might expect a site to operate within a broad range subject to access, design, environmental and infrastructure constraints. A smaller parcel with heritage buildings may sit toward the lower end. A large unconstrained site with excellent access may sit toward the upper end.

The range should not become a hidden invitation to negotiate without evidence. The criteria for moving within it should be explicit.

A design-led density approach begins with capacity, not a target pasted onto the site

London’s current planning framework is an important example of a shift away from relying on a rigid density matrix as the single determinant of site capacity. Its design-led approach asks how much development a site can appropriately accommodate while meeting policy and design requirements.

The idea is not that density no longer matters.

It means the density number is an output of a tested design and site context rather than a substitute for design.

That distinction is useful anywhere. A plan can set strategic expectations while still requiring the actual building envelope, public realm, access, open space and environmental performance to work.

Transit proximity can justify an intensity expectation without making transit the owner of density

Many planning systems expect more homes or floor area in places with strong public-transport access.

That is a location-allocation decision.

This article does not own transport planning. It only clarifies how the resulting residential intensity is measured.

If a policy says “at least seventy dwellings per hectare near a station,” the density analyst still needs to know whether that is net or gross, which parcels fall inside the station area, how mixed-use land is treated and whether the minimum applies parcel by parcel or across a master-planned area.

A location policy without a measurement convention becomes difficult to administer consistently.

Mixed-use development makes the denominator more difficult

Imagine a one-hectare site with shops on the ground floor, offices on two levels and 150 apartments above.

Is the residential density 150 dwellings per hectare?

That may be a perfectly useful site-density measure because all 150 dwellings occupy a development whose site is one hectare.

But if a regional housing study wants to compare purely residential land consumption, counting the full mixed-use hectare as residential land may not match its purpose.

Some studies allocate land proportionally. Others use the whole site. Others report residential floor area, dwellings and jobs separately.

There is no useful universal answer detached from the decision. The methodology should be declared and used consistently.

Vertical mixed use creates another counting problem

When residential and non-residential uses share the same land vertically, the site cannot be divided neatly on a two-dimensional map.

This is one reason floor-area measures remain useful alongside dwelling density. FAR can describe total built intensity while dwelling density describes the residential-unit count.

A sophisticated planning dashboard may therefore report several measures together rather than forcing one number to carry every meaning.

Unit size can change density without changing the building envelope

Suppose an apartment building contains 12,000 square metres of residential floor area inside a fixed approved envelope.

A design with larger family units might contain 100 dwellings.

A design with smaller studios and one-bedroom homes might contain 180.

The external building could be almost unchanged.

Dwelling density rises because the numerator changes.

This can affect lift demand, refuse, bicycle storage, mail, deliveries, water fixtures, parking and household profile even though FAR and height remain constant.

Planning systems that regulate only floor area should understand this unit-count effect where infrastructure or amenity is sensitive to the number of households.

Bedroom density can sometimes reveal more than dwelling density

One hundred studio apartments and one hundred four-bedroom homes both count as one hundred dwellings.

The potential resident capacity is very different.

For some infrastructure studies, bedrooms per hectare or estimated residents per hectare can therefore supplement dwelling density.

That does not mean bedrooms are a universally better metric. It means the numerator should match the question.

Occupancy caps and planning density should remain separate concepts

A jurisdiction may regulate how many people can occupy a dwelling for health, safety or tenancy reasons.

That is not the same as residential density planning.

One is a rule about occupancy inside a unit. The other is a measure of homes or people over land.

Keeping them separate prevents a density debate from becoming an indirect and often inaccurate debate about household behaviour.

Density can be averaged across a master-planned site

A large development rarely uses one uniform intensity.

Homes may be concentrated near a centre, with lower-rise areas toward an edge. One parcel may contain a park. Another may hold a community facility. A third may contain a taller mixed-use building.

If the planning permission establishes an average density across the whole master plan, individual subplots can vary provided the overall calculation remains within the approved framework.

The approval should state whether density is:

fixed on every parcel;
averaged across a phase;
averaged across the entire plan;
transferable between subplots;
or subject to minimum and maximum parcel ranges within an overall total.

Otherwise later phases can discover that early phases quietly consumed the capacity.

Phasing needs a density ledger

Suppose a master plan is approved for 2,000 dwellings across twenty hectares of net residential land.

Early phases deliver 1,200 dwellings on ten hectares.

The remaining ten hectares now need to carry 800 dwellings if the total is fixed, producing a lower average in the later phases. If early phases underdeliver, the later phases may need greater intensity.

A density ledger records approved land area, delivered units, remaining units and any transfers between parcels.

This is especially important when infrastructure or affordable-housing commitments are tied to the total dwelling count.

Rounding can create real units

Consider a 0.37-hectare parcel subject to a maximum of forty dwellings per hectare.

The mathematical result is 14.8 dwellings.

Does the site permit fourteen units, fifteen units, or does the density operate only as a guide rather than a hard numerical cap?

A code should specify the rounding convention.

Across hundreds of small sites, rounding can add or remove a meaningful amount of housing capacity.

Very small sites can make density ratios look extreme

One additional dwelling on a tiny parcel can cause a large percentage change in dwellings per hectare.

A 0.05-hectare lot with one dwelling equals twenty dwellings per hectare.

Add a second dwelling and the density becomes forty dwellings per hectare.

The physical change may be one modest accessory home.

This is why density caps can behave awkwardly for small-lot infill. Form, floor area, setbacks or unit-type rules may be more direct controls depending on the planning objective.

Minimum lot size and density can pull in opposite directions

If each dwelling requires a large minimum lot, there is a structural ceiling on achievable dwelling density.

For detached housing, a one-hectare site cannot contain sixty one-unit lots if every lot must be at least 300 square metres before roads and other land are considered.

This is why zoning reforms should test all dimensional controls together. A policy can announce a higher density while a minimum lot-size rule quietly prevents it.

The Minimum Lot Size owns that parcel constraint.

Setbacks can make a nominal density physically impossible

A density target may permit eighty dwellings per hectare.

If deep front, rear and side setbacks leave a tiny building footprint, the permitted density may exist only on paper.

That is why density should be tested against the buildable envelope described in The Setback Line.

The effective development capacity is the intersection of all applicable controls.

Height can make a density target achievable—or impossible

When land coverage is constrained, additional dwellings often require additional floors.

A low height cap can therefore limit the achievable unit count even when the density policy permits more.

Conversely, very generous height does not guarantee high dwelling density if the approved floor area is low or the units are large.

This is why The Building Height Rule, FAR, coverage and density should be modelled together rather than read as independent lines in a code.

Density bonuses should be kept separate from baseline density

A site might have a baseline maximum of eighty dwellings per hectare and qualify for additional capacity if it provides a specified public benefit.

The bonus calculation should show:

baseline density;
qualifying bonus percentage or unit count;
maximum post-bonus density;
the public benefit that earns the bonus;
and any separate FAR or height changes needed to make the extra homes physically achievable.

The Incentive Zoning article owns the exchange mechanism. The density measure makes the arithmetic transparent.

Density transfers need a sending and receiving ledger

Some plans allow housing capacity to move between parcels or development phases.

If one parcel contains a protected habitat, heritage building or public open space, its expected units may be transferred to a more buildable parcel.

The system should record the original capacity, transferred amount, receiving capacity and remaining entitlement so the same density is not counted twice.

Where development rights themselves become transferable instruments, The Development Rights Market owns the larger mechanism.

Density at a station, centre or corridor should specify the geography

Policies often describe “station-area density” or “corridor density.”

But the measured result changes depending on whether the area is a 400-metre radius, an 800-metre walking catchment, a set of designated parcels or a political boundary.

A circular radius can include land that is physically close but inaccessible because of a river, motorway or railway. A walking catchment can better represent actual access but is more complex to calculate.

Again, measurement should follow the planning question.

Density gradients are more informative than one district average

A planned district can intentionally place higher residential intensity near its strongest access and mixed-use centre, then step down toward sensitive edges.

The district average may be forty-five dwellings per hectare, while individual residential parcels range from twenty-five to 120.

The gradient explains the spatial strategy better than the average alone.

Maps, ranges and parcel schedules should therefore accompany a master-plan density target where distribution matters.

Gross density makes public land visible

Net density is useful for comparing the efficiency of residential development parcels.

Gross density reminds us that neighbourhoods also require streets, public spaces, drainage, utilities and civic land.

If two new communities each achieve sixty net dwellings per hectare but one requires far more land for roads and fragmented leftover space, their gross densities can differ substantially.

Neither number is automatically “correct.” Together they reveal different parts of land consumption.

Net density can hide the infrastructure footprint if used alone

Suppose a residential superblock contains 300 dwellings on five net residential hectares: sixty dwellings per hectare.

If reaching and serving that superblock requires three additional hectares of roads, interchanges and engineered drainage outside the net boundary, the net number tells only part of the spatial story.

Gross density can expose the wider land requirement.

This is particularly important when comparing development patterns rather than merely checking one parcel against zoning.

Density can support infrastructure planning without claiming ownership of infrastructure

More dwellings and more people can change demand on roads, transit, schools, parks, water, sewers, health services and community facilities.

Those systems have their own owners elsewhere in the eduKate town-planning library.

The density measure contributes one input: how much residential activity is expected within a defined area.

Infrastructure models then need their own assumptions about trip generation, age profile, peak demand, consumption, service standards and network capacity.

One density figure should not be asked to predict all of those outcomes by itself.

A dwelling is not a standard unit of demand

A studio, a three-bedroom apartment and a five-bedroom house all count as one dwelling.

For zoning simplicity that may be exactly what is needed.

For detailed infrastructure planning, planners may need a second layer: bedroom mix, occupancy assumptions, resident age profile or floor area.

The lesson is not to abandon dwelling density. It is to avoid treating a simple land-use metric as a complete demand model.

Jobs density and residential density should not be added casually

A mixed-use district can contain 10,000 residents and 15,000 jobs.

Adding them to produce “25,000 people” can be useful for some activity-density analyses.

It can also double-count people who both live and work in the area, and it obscures different temporal patterns. Residents are present overnight. Workers peak during business hours. Visitors add another layer.

Activity density is therefore a separate analytical tool and should state whether it counts residents, jobs, students, visitors or some combination.

Daytime and night-time density can be different cities

A central business district may be extremely busy at midday and quiet at 2 a.m.

A residential district can show the opposite pattern.

Static dwelling density cannot describe this temporal change.

For emergency planning, transport operations or public-space management, time-specific population or activity estimates may be needed.

This is another reminder that density is a family of measurements, not one universal statistic.

Density is a ratio, so every mapping choice matters

Suppose a neighbourhood boundary follows a river on one side.

If the water surface is included in the municipal area, population density falls. If only developable land is counted, it rises.

Suppose an airport occupies one quarter of a city’s land. Citywide population density drops even though no residential street became less intense.

That is why maps should accompany density claims whenever the boundary is not obvious.

Density comparisons should normalise both the unit and the geography

Before comparing two places, check:

Are both numbers dwellings, people or floor area?
Are both per hectare, per acre or per square kilometre?
Are both gross or both net?
Do both geographies include streets?
Do both include parks and schools?
Is one a development parcel and the other an administrative district?
Are the values measured at the same date?
Are vacant units treated consistently?

Only after those questions are answered does the comparison become analytically useful.

A density heat map can reveal distribution that an average hides

Instead of assigning one value to an entire district, planners can calculate density for smaller grid cells, blocks or parcels and map the result.

The pattern may reveal a concentrated centre, corridors, low-density gaps, abrupt transitions or clusters of underused land.

But cell size matters. A coarse grid smooths differences. A fine grid can become noisy.

The map should therefore state the spatial unit used to calculate the density.

Density targets should be tested against the parcel inventory

A city may announce that a district should average sixty dwellings per hectare.

The next question is whether the inherited parcels can actually deliver that level under current lot sizes, setbacks, heights, coverage, heritage constraints and existing buildings.

A parcel-level capacity model can estimate how many dwellings fit under realistic assumptions, then aggregate those results to the district.

This turns a policy aspiration into a testable land-capacity proposition.

Existing density and permitted density should not be confused

A neighbourhood may currently contain twenty dwellings per hectare while zoning permits sixty.

The existing density describes what is there now.

The permitted density describes a regulatory ceiling or range.

The realistic future density may still be much lower because redevelopment happens gradually, parcels are fragmented, buildings have long economic lives and owners make independent decisions.

Planning forecasts should therefore distinguish existing, permitted and likely delivered density.

The zoning maximum is not a forecast

This distinction prevents a common modelling error.

If every parcel is assumed to redevelop immediately to the maximum density, infrastructure forecasts can become unrealistic.

More credible scenarios apply redevelopment probability, market feasibility, site assembly, existing building age, ownership, environmental constraints and phasing.

The density control defines legal capacity. The forecast estimates behaviour within that capacity.

The urban growth boundary changes the denominator at a metropolitan scale

When a region limits outward expansion, additional population and housing have to be accommodated through some combination of redevelopment, infill, smaller lots and greater intensity inside the boundary.

That does not dictate one local density everywhere.

It changes the metropolitan land budget.

The Urban Growth Boundary owns that regional mechanism. The density measure helps show how much housing the land inside the boundary can support under different assumptions.

A worked example: sixty homes on one hectare

Take one hectare of net residential land.

Place sixty homes on it.

The net dwelling density is sixty dwellings per hectare.

Now imagine the sixty homes as three different schemes.

Scheme A: sixty narrow rowhouses averaging one home per lot.
Scheme B: six apartment buildings with ten homes each.
Scheme C: one taller apartment building with sixty homes and a large shared garden.

All three have the same dwelling density.

They do not have the same height, coverage, open-space pattern, construction cost, street frontage, accessibility or household mix.

The density number answers one question and deliberately leaves the others open.

A second worked example: the ten-hectare neighbourhood

Return to the ten-hectare neighbourhood with 360 homes and six net residential hectares.

The gross density is thirty-six dwellings per hectare.

The net density is sixty.

Suppose a second neighbourhood also contains 360 homes but uses seven net residential hectares and only three hectares for all other land.

Its gross density is still thirty-six dwellings per hectare because the total area remains ten hectares.

Its net residential density is about 51.4 dwellings per hectare.

The same gross density can therefore coexist with different net densities because the internal allocation of land differs.

This is why both measures can be informative.

A third worked example: the mixed-use tower

A one-hectare site contains a retail podium, offices and 200 apartments.

If the whole site is used as the denominator, residential site density is 200 dwellings per hectare.

The building also contains 25,000 square metres of non-residential floor area and 30,000 square metres of residential floor area.

One number cannot describe all of that intensity.

A useful planning schedule might therefore report:

200 dwellings;
200 dwellings per hectare;
30,000 square metres residential floor area;
25,000 square metres non-residential floor area;
total FAR or plot ratio;
and an estimated residential population based on a stated occupancy assumption.

Each measure answers a different planning question.

A fourth worked example: the density target that cannot fit

A 0.8-hectare parcel is zoned for a target of 100 dwellings per hectare.

That implies eighty dwellings if the target applies directly to net site area.

But setbacks, tree protection and access leave only thirty per cent of the site for building footprint. Height is capped at three storeys. After circulation and common areas, the realistic design holds only forty-eight homes.

The density target says eighty.

The physical envelope says forty-eight.

The planning authority now has a useful diagnostic question: which rule is intentional, and which is contradicting the adopted objective?

Without the calculation, the contradiction remains hidden inside separate tables.

Density monitoring should report what was approved and what was built

If a plan expects higher residential intensity, the city should track delivery.

Useful fields include:

site area and density denominator;
approved dwelling count;
completed dwelling count;
approved and completed density;
unit mix;
estimated or observed population;
gross and net land take where relevant;
and any bonus or transferred units.

Over time, this shows whether the code produces the density the plan expects or whether other controls and market conditions systematically reduce it.

Repeated underdelivery is planning information

If every site in a district is zoned for eighty dwellings per hectare but completed projects average forty-five, the city should not simply accuse every applicant of lacking ambition.

The pattern may reveal:

height limits that suppress usable floor area;
setbacks that consume narrow lots;
parking requirements that dominate the ground plane;
unit-size requirements;
market demand for larger homes;
infrastructure constraints;
heritage or environmental restrictions;
or land assembly problems.

Density monitoring therefore becomes a diagnostic for the wider planning system.

Repeated overdelivery can be information too

If nearly every project seeks and receives relief above the stated maximum, the formal density cap may no longer represent the city’s actual policy.

The city should examine whether the mapped standard should be updated openly rather than maintained as a nominal limit followed by routine exceptions.

The same principle appears throughout good zoning administration: repeated exceptions are feedback about the rule.

Density should be published with metadata

A modern density dataset should not contain only a number.

It should record:

numerator definition;
denominator definition;
unit of area;
geographic boundary;
measurement date;
whether the value is existing, approved or forecast;
whether vacant units are included;
and the source of the underlying dwelling or population count.

That metadata turns a statistic into reusable planning evidence.

The Density Measure audit

Numerator: Are we counting dwellings, people, bedrooms, jobs or floor area?
Denominator: What exact land is included?
Unit: Is the result per hectare, acre, square kilometre or another unit?
Gross or net: Is the convention stated explicitly?
Roads: Are local and major streets included in the denominator?
Parks: Are public open spaces included?
Schools: Are institutional sites included?
Utilities: Are drainage and infrastructure reserves included?
Mixed use: How is land shared vertically between residential and non-residential uses treated?
Scale: Is this parcel, block, neighbourhood, district or city density?
Boundary: Is the calculation geography mapped?
Date: When was the dwelling or population count measured?
Vacancy: Are empty homes included in dwelling density?
Population: Is resident count observed or estimated?
Household size: What occupancy assumption is used?
Unit mix: Do studios and family homes count equally?
Bedrooms: Would bedroom density improve a specific service-demand analysis?
Existing: Does the number describe what is currently built?
Permitted: Does it describe legal maximum or minimum capacity?
Forecast: Does it describe a likely future rather than a zoning ceiling?
Minimum density: Is there a lower intensity below which land is considered underused?
Maximum density: Is there an upper limit?
Range: Are the reasons for moving within a density range stated?
Rounding: How are fractional units handled on small sites?
Master plan: Is density fixed per parcel or averaged across a larger area?
Phasing: Are delivered and remaining units tracked?
Transfers: Can units move between parcels, and is there a ledger?
Bonus: Is bonus density separated from baseline capacity?
FAR: Can the permitted floor area support the intended dwelling count?
Height: Can the units physically fit inside the vertical envelope?
Coverage: Does the ground footprint permit the intended building type?
Setbacks: Does the horizontal envelope make the target achievable?
Minimum lot size: Does subdivision policy contradict the density objective?
Parking: Does parking consume the land needed to achieve the intended intensity?
Quality: Are housing quality standards tested directly rather than inferred from density?
Crowding: Is interior occupancy distinguished from urban density?
Infrastructure: Is density used as one input rather than a complete demand model?
Comparison: Have units and geographies been normalised before comparing places?
Monitoring: Are approved and completed densities tracked?
Feedback: Does repeated under- or overdelivery trigger a review of the code?

The most important word in a density argument is often the word nobody says

Net.

Gross.

Residential.

Population.

Existing.

Permitted.

Forecast.

Each missing qualifier can change what the number means.

A city does not become well planned because it reaches the highest possible dwellings-per-hectare figure. Nor does low density automatically create quality. Density is a measurement tool. Its value comes from using the right numerator, the right denominator and the right scale for the decision.

The strongest density policy does not use one number to stand in for height, housing quality, congestion, infrastructure or neighbourhood character. It defines the ratio precisely, tests whether the physical zoning can deliver it, maps how it varies across space, and then lets the other planning systems do their own jobs.

Sources and further reading

Continue reading: Planning rules, permissions and land rights · Full Town Planning Series Index · Urban Planning Master Edition.

How Town Planning Works | TPW-0158 — The Density Measure: How Gross and Net Density, Dwellings per Hectare and People per Area Change What “High Density” Means