New apartment buildings tend to open with quartz counters, stainless appliances and a rent that most people in the neighborhood cannot pay. The obvious question is why nobody builds the plain version. Take out the quartz, put in laminate, skip the rooftop deck, and rent the result for less.
The answer is that a developer does not set the rent the way a shopkeeper sets a price. The rent a new building needs is the sum of what its capital and its operations cost, and the market decides whether to pay it. If the market will not, the building is not built. "Cheaper apartments" therefore has to mean one of three things: less capital per unit, a lower cost of that capital, or a lower cost to operate the building once it opens. Finishes fall into the first category, and they are a small part of it.
This article walks through the cost of an illustrative 120-unit building and then tests eight levers, one at a time, to see how much each moves the rent the building requires. The numbers are illustrative assumptions, not measurements of any real market, and every one of them is listed below.
What the rent has to cover
The building is Reference Project B, the podium project, one of the two illustrative buildings used across this site: a 120-unit mid-rise over a structured garage, with a total development cost of $30,000,000, or $250,000 per unit. That figure is the sum of six pieces. Land is $3,600,000. Hard costs, the physical construction, are $19,500,000, and that figure includes 120 structured parking spaces at $30,000 each, or $3,600,000, which leaves $15,900,000 for the building itself. Soft costs such as design, engineering, legal work, insurance and marketing are $3,300,000, of which $1,440,000, or $12,000 per unit, is permit and impact fees. The developer fee is $1,200,000. Financing costs during construction are $1,800,000, and lease-up and operating reserves are $600,000.
Inside that total, we assume the interior finishes a tenant actually sees and touches, meaning flooring, cabinets, counters, fixtures and appliances, come to $20,000 per unit, or $2,400,000 in all. That is 8% of the project. The other 92% is land, structure, parking, systems, fees, professional work, financing and reserves. A tenant walking through a model unit forms an impression of the building from the 8%.
The $30,000,000 has to be paid for. We assume a lender provides 65% of cost as a permanent loan, $19,500,000, at 6.5% interest amortized over 30 years, which produces annual debt service of $1,479,039. The remaining 35%, $10,500,000, is equity, and we assume its investors require an 8% annual cash yield, or $840,000. Together those two claims mean the building must produce net operating income of $2,319,039 a year before anyone has been paid a dollar of profit beyond that yield. Add operating expenses, which we assume at $8,200 per unit per year, or $984,000, and gross up for a 5% vacancy allowance, and the building needs $3,476,883 of gross revenue a year. Divided by 120 units and twelve months, that is a required rent of $2,415 per unit per month.
It helps to see the $2,415 in three pieces, each already grossed up for vacancy: $1,081 covers debt service, $614 covers the equity return, and $719 covers operating expenses. About seventy percent of the rent is the cost of the capital that built the building.
This is what practitioners mean by a minimum feasible rent. The lender sizes its loan to the debt-service coverage it requires, the equity requires its yield, and the rent that satisfies both is a floor. Above the floor, the developer has a project. Below it, the lender will not lend, the investors will not invest, and nothing is built. The developer does not get to pick a lower number out of goodwill any more than a tenant gets to pick a lower rent by asking. A line-by-line walk through the same arithmetic is in why a new apartment costs $2,000 a month, which runs it on Reference Project A, the site's cheaper garden building with surface parking, and arrives at $1,996 rather than $2,415; the sequence of decisions that leads up to either number is in how housing gets built.
Why cheaper finishes barely register
Suppose the developer takes the obvious advice and cuts the finish budget by a quarter, from $20,000 to $15,000 per unit. Laminate instead of stone, a cheaper appliance package, vinyl plank instead of engineered wood. Across 120 units that saves $600,000, which is 2% of the project.
Cut interior finishes by 25%
−$600,000
−$34/month
The rent effect follows from the same financing structure as the building as a whole. Of the $600,000 no longer spent, 65% would have been borrowed and 35% would have come from equity. The borrowed part no longer needs debt service; the equity part no longer needs its 8% yield.
Debt avoided = $600,000 × 65% = $390,000; equity avoided = $600,000 × 35% = $210,000
Debt service avoided = $390,000 × 7.585% mortgage constant = $29,582 per year
Equity return avoided = $210,000 × 8% = $16,800 per year
Required NOI reduced by $29,582 + $16,800 = $46,382 per year
Grossed up for 5% vacancy: $46,382 ÷ 0.95 = $48,823 of revenue
$48,823 ÷ 120 units ÷ 12 months = $34 per unit per monthThe 7.585% figure is the annual mortgage constant at 6.5% interest over 30 years.
Thirty-four dollars is 1.4% of the $2,415 rent. It is not nothing, but it is not the difference between a building people can afford and one they cannot. And the $34 is a ceiling on the benefit rather than an estimate of it. Cheaper finishes can also lower the rent the market is willing to pay for the unit, and cheaper materials tend to wear faster and cost more to repair and replace at turnover. Both of those effects push the other way, so the net gain from a plainer unit can be smaller than $34, and it can be negative. This article does not model the revenue side; it simply notes that the cost side is small.
That is the core of the answer to the question in the title. The apartment the tenant sees is a thin layer on top of a large, mostly invisible cost base, and the visible layer is not where the money is.
The levers that actually move the number
If finishes are a small lever, the useful question is which levers are large. The rest of this section applies the same arithmetic to seven other changes, each computed alone with everything else held at the base case.
Land and density
Land is $3,600,000 of the base project, or $30,000 per unit. If we assume the same site could be bought 25% cheaper, saving $900,000, the required rent falls by $51 per month, or 2.1%. A developer rarely controls that number directly; land sells for what the next buyer will pay, and the next buyer is usually running the same arithmetic.
What a developer sometimes can control, subject to zoning, is how many units the land carries. Density does not make land cheaper; it spreads the same land cost over more units. We assume the same site could hold 150 units instead of 120, an increase of 25%. Land stays at $3,600,000. Hard costs, the developer fee, financing costs and reserves scale with the unit count. Of the soft costs, we assume half are fixed regardless of size and half scale with it. The total becomes $36,187,500, which is $241,250 per unit against $250,000 in the base case, a saving of $8,750 per unit. The required rent falls to $2,355, a change of $59 per month, or 2.5% (the two rents and the difference are each rounded from unrounded figures, which is why they do not subtract exactly).
The density result depends heavily on what land costs. Under our assumptions land is 12% of the project, so spreading it thinner has a modest effect. If the same site cost $9,000,000 instead, with everything else unchanged, the same move from 120 to 150 units would lower required rent by $120 per month. Land is the input that varies most from one place to another, which is why the same zoning change can be a small matter in one market and a large one in another.
Parking
The base project provides one structured parking space per unit, 120 spaces at $30,000 each, which is $3,600,000 of the hard costs. We assume the parking ratio could be cut from 1.0 to 0.5 spaces per unit, removing 60 spaces and $1,800,000 of cost. The required rent falls by $102 per month, or 4.2%.
That single change is three times the effect of cutting finishes, and it comes from a part of the building that most tenants never think about as part of their rent. Whether it is available depends on the site and the rules: in many places the parking ratio is set by the zoning code rather than by the developer, and in places where tenants expect a space, providing fewer can lower the rent the building achieves. The parking arithmetic is worked through in detail in what a parking space adds to rent.
Construction type
The largest construction lever in the model is not what goes inside the walls but what the walls are made of. Buildings of different heights are built with different structural systems, and the lower-rise systems, such as wood frame, generally cost less per square foot than concrete or steel. We assume a change of construction type that reduces the non-parking hard costs, the $15,900,000 for the building itself, by 15%, or $2,385,000. The required rent falls by $135 per month, or 5.6%.
The tradeoff is that lower-cost construction types are usually height-limited. A building that cannot go as tall needs more land per unit for the same number of apartments, which is the opposite of the density lever. On a cheap site the wood-frame building wins; on an expensive one the taller, costlier structure may still produce the lower rent per unit because it spreads the land further. The two levers interact, and a developer choosing between them is really choosing based on what the land costs.
Fees
Permit and impact fees are $1,440,000 of the base project, $12,000 per unit. We assume they could be halved, to $6,000 per unit, saving $720,000. The required rent falls by $41 per month, or 1.7%. Fees are set by the jurisdiction rather than by the developer, and they pay for things such as sewer capacity, parks and plan review that someone has to pay for. The question of who is a policy question; the arithmetic only says what the fee costs the tenant.
Time
A project that waits twelve months longer for its approvals carries its land and its early spending for twelve months longer, and its construction prices move over that period. In what a year of development delay costs, we assume a year of delay adds $1,265,200 of carrying costs and escalation to the base project. Removing that year, meaning approvals twelve months faster, saves the same amount and lowers the required rent by $71 per month, or 3.0%. The entitlement process is not a line item on the building, but it shows up in the rent as if it were.
The price of money
The last lever is not a construction choice at all. We assume the permanent loan rate falls one percentage point, from 6.5% to 5.5%. Annual debt service on the same $19,500,000 loan falls from $1,479,039 to $1,328,626. Grossed up for vacancy and divided across the units, the required rent falls by $110 per month, or 4.6%, and no one has changed a single drawing.
That $110 is deliberately narrow, and it understates the lever. It changes the permanent loan only. The $1,800,000 of construction-period financing in the budget is held at the base case, even though a construction loan priced a point lower would have accrued roughly $195,000 less interest over the build, which is another $11 per month. Holding it fixed keeps this lever computed the same way as the other seven, each of which changes one budget line and nothing else; the full effect of a rate move, including construction interest and the equity yield repricing alongside it, is worked through in what happens when interest rates rise 1 percent, on the site's garden reference project.
A developer has no more control over the interest rate than a tenant does, but its effect on the rent is larger than any single design choice except construction type.
| Lever | Change assumed | Capital change | Rent change per month | Share of base rent |
|---|---|---|---|---|
| Cheaper finishes | Finish budget cut 25% ($20,000 to $15,000 per unit) | −$600,000 | −$34 | −1.4% |
| Lower fees | Permit and impact fees halved ($12,000 to $6,000 per unit) | −$720,000 | −$41 | −1.7% |
| Cheaper land | Land 25% cheaper | −$900,000 | −$51 | −2.1% |
| Density +25% | 150 units on the same site | −$8,750 per unit | −$59 | −2.5% |
| Faster approvals | Twelve months less carrying cost and escalation | −$1,265,200 | −$71 | −3.0% |
| Half the parking | 1.0 to 0.5 structured spaces per unit (60 fewer) | −$1,800,000 | −$102 | −4.2% |
| Wood-frame construction | Non-parking hard costs cut 15% | −$2,385,000 | −$135 | −5.6% |
| Interest rate −1 point | Permanent loan 6.5% to 5.5% | n/a (financing) | −$110 | −4.6% |
Rent impact of each lever
See the numbers
| Category ($ per unit per month) | Change in required rent |
|---|---|
| Cheaper finishes | −$34 |
| Lower fees | −$41 |
| Cheaper land | −$51 |
| Density +25% | −$59 |
| Faster approvals | −$71 |
| Half the parking | −$102 |
| Wood-frame construction | −$135 |
| Interest rate −1 point | −$110 |
Why the levers are not simply additive
Adding up the seven capital levers, and leaving out the interest rate because it is not a choice anyone on the project makes, gives a combined reduction of roughly $493 per month, which would put the required rent near $1,922. That number is worth having as a rough outer bound, but it is not a plan, for three reasons.
- Several levers conflict. Density pushes toward a taller building; wood frame pushes toward a shorter one. A project cannot fully take both.
- Several levers are not available everywhere. Halving parking depends on the site, the transit around it and what the zoning code allows. Faster approvals depend on the jurisdiction. Cheaper land depends on where the land is.
- Each lever was computed with everything else held constant. Once one changes, the base the others act on has changed too, so the effects do not stack exactly.
There is also a lever on the other side of the ledger. Of the $2,415 rent, $719 covers operating expenses, and of the assumed $8,200 per unit per year in operating expenses, $2,500 is property tax. A tax abatement is a distinct instrument that works on the operating line rather than the capital line; it is examined separately in could government finance housing more cheaply and not computed here.
What the table shows, taken as a whole, is a ranking. Of the eight levers, the one a developer controls most completely, finishes, is the smallest. The four largest are construction type, interest rates, parking and approval time, and each of those is set mostly or entirely by something other than the developer: building codes, the bond market, the zoning code and the permitting process.
Reference Project B (the podium project): an illustrative 120-unit mid-rise over a structured garage of 120 spaces; total development cost $30,000,000 ($250,000 per unit); operating expenses $8,200 per unit per year; required rent $2,415 per unit per month. Why this project: a building with structured parking and a costlier structural system has the widest set of levers left to pull, which is the subject of this article. Both reference projects are defined side by side on our Methodology page.
We assume land of $3,600,000; hard costs of $19,500,000, including 120 structured parking spaces at $30,000 each ($3,600,000), leaving $15,900,000 for the building itself; soft costs of $3,300,000, of which permit and impact fees are $1,440,000 ($12,000 per unit); a developer fee of $1,200,000; construction-period financing costs of $1,800,000; and lease-up and operating reserves of $600,000.
We assume interior finishes (flooring, cabinets, counters, fixtures, appliances) of $20,000 per unit, $2,400,000 in total, or 8% of total development cost.
We assume financing of 65% debt ($19,500,000) at 6.5% interest amortized over 30 years, giving a mortgage constant of 7.585% and annual debt service of $1,479,039, and 35% equity ($10,500,000) requiring an 8% cash yield, or $840,000 per year.
We assume operating expenses of $8,200 per unit per year ($984,000): property taxes $2,500 per unit (1.0% of development cost), insurance $1,000, operations $3,500, owner-paid utilities $800 and capital reserves $400, the last four above the garden project's because elevators, a podium structure and an enclosed garage have to be run, lit, ventilated and eventually replaced. Vacancy allowance 5%. Base required rent is $2,415 per unit per month: $1,081 for debt service, $614 for the equity return and $719 for operating expenses, each grossed up for vacancy.
For the finishes lever we assume a 25% cut, from $20,000 to $15,000 per unit, saving $600,000. Effects on achievable rent, turnover and repair costs are not modeled.
For the fees lever we assume permit and impact fees halved, from $12,000 to $6,000 per unit, saving $720,000.
For the land lever we assume land 25% cheaper, saving $900,000.
For the density lever we assume 150 units on the same site. Land is unchanged at $3,600,000; hard costs, developer fee, financing costs and reserves scale in proportion to units; half of soft costs are fixed and half scale with units. Total cost becomes $36,187,500, or $241,250 per unit. For the comparison in the text we assume an alternative site costing $9,000,000 with all other assumptions unchanged.
For the approvals lever we assume twelve months of avoided carrying costs and escalation of $1,265,200, the figure derived in the delay article.
For the parking lever we assume the ratio falls from 1.0 to 0.5 structured spaces per unit, removing 60 spaces at $30,000 each, or $1,800,000.
For the construction type lever we assume a change that reduces non-parking hard costs by 15%, or $2,385,000, with no change to land or unit count.
For the interest rate lever we assume the permanent loan rate falls from 6.5% to 5.5%, reducing annual debt service from $1,479,039 to $1,328,626, with construction-period financing costs held at $1,800,000 and the equity yield held at 8%. Both are held fixed so that this lever is computed like the other seven, one budget line at a time; the figure is therefore a floor. A construction loan priced a point lower would have accrued roughly $195,000 less interest, worth about $11 more per month, and equity that repriced with rates would push in the other direction.
All rent figures use the site-wide rent math. Annual debt service is the mortgage payment on the capital cost multiplied by the loan-to-cost ratio, at the stated interest rate over the amortization period, times twelve. Annual equity return is the capital cost multiplied by one minus the loan-to-cost ratio, multiplied by the target equity yield. Required net operating income is the sum of the two. Required revenue is required net operating income plus operating expenses, divided by one minus the vacancy rate. Required rent per unit per month is required revenue divided by units and by twelve. Nothing is rounded until the final figure.
Each lever is computed alone, holding every other assumption at the base case, and its rent effect is the difference between the base required rent and the required rent with that one change. Interactions between levers are ignored, which is why the combined figure of roughly $493 per month is an outer bound rather than a forecast.
For capital levers, the rent effect equals the capital change multiplied by the loan-to-cost ratio and the mortgage constant, plus the capital change multiplied by the equity share and the equity yield, grossed up for vacancy and divided across 120 units and twelve months. The density lever is computed differently: total cost is rebuilt at 150 units with land fixed, hard costs, developer fee, financing and reserves scaled by 1.25, and soft costs split half fixed and half scaled, then the required rent is recomputed per unit. The interest rate lever changes only the debt service on the permanent loan.
Revenue-side effects of finishes, including any change in the rent tenants will pay or in turnover and repair costs, are not modeled. Effects of parking on achievable rent are likewise not modeled. Property tax abatement is not computed here.
The budget this article takes apart is a model. Its cost splits, its finish allowance, its fee schedule and its construction-type differential are assumptions we set out in order to show which levers are large and which are small, not measurements taken from a cost survey or from a project that was actually built. Each of those inputs is local: fees are set by a municipality, construction differentials by local labor and code, parking by the site. A reader who wants to know which lever is biggest in a particular city should price these there and rerun the same eight comparisons.
A market with expensive land. If land were a larger share of cost, the density lever would grow (to about $120 per month at $9,000,000 of land under our assumptions) and the case for taller, costlier construction would strengthen, since the two levers trade against each other.
Surface rather than structured parking. If parking were built at grade on a cheap site, the cost per space would be far lower and the parking lever would shrink, though the land per unit would rise.
A building already at the cheapest construction type. A low-rise wood-frame project has no construction-type lever left to pull; its remaining levers are land, parking, fees and time.
A lower interest rate. Every capital lever is smaller in dollar terms when money is cheaper, because each avoided dollar of cost carries less debt service. The ranking of levers would not change, but their size would.
Rents restricted by program. If a building's rents are capped by a subsidy program, the market no longer decides whether to pay the required rent; a gap between the capped rent and the required rent has to be closed by a subsidy, and the levers here determine how large that gap is.
Labor and materials prices. A rise in construction prices raises the non-parking hard costs and enlarges every lever that acts on them; a fall does the reverse. The finishes share would move only if finish prices moved differently from the rest of construction.
What this means
The question "why can't developers just build cheaper apartments" assumes the apartment is where the cost lives. Mostly it is not. The finishes a tenant sees are 8% of the project, and cutting them by a quarter moves the required rent by about $34 a month, before any loss of rent or extra wear from cheaper materials. The choices that would make apartments meaningfully cheaper are mostly not about the apartment. They are about land and how many homes it carries, about rules that set parking ratios and building heights and fees, about how long approvals take, and about the price of money.
A developer working inside a fixed set of those rules has a narrow range. Within it, the developer will generally build the version the market will pay the most for, because the required rent is a floor and the margin above it is thin. Changing the range means changing the rules, the land, the timeline or the rate, and each of those belongs to someone other than the developer.
None of this says which rules should change. Parking ratios, height limits and impact fees exist for reasons, and the people who live near a proposed building bear some of the consequences of relaxing them. What the reader now has is a rough scale: roughly $34 for finishes, $41 for fees, $51 for land, $59 for density, $71 for a year of time, $102 for parking, $135 for building type and $110 for a point of interest, under one illustrative set of assumptions. Any argument about cheaper apartments should be able to say which of those it is talking about.