Housing Development Dashboard

Development Calculator

Top 6 Factors Influencing New Development

We believe that providing a transparent, open resource for policymakers can improve local decision making. The Housing Development Dashboard shows how local policies and development factors impact the odds that a housing development gets built. The Development Calculator focuses on the most important factors supported by the literature and local development experts. The methodology and default assumptions were initially vetted through conversations with area development experts, data collection, and analysis from January to May of 2016. Key model assumptions were updated in September of 2017 to reflect changing market conditions.

The calculator works best for properties of 50 units or more, projects in which the developer has not yet entered into an option agreement, and where the land seller is motivated to sell. In reality, many of these factors move together, so users should be careful to interpret results significantly different from existing market conditions. The tool is currently in βeta testing.

Market Factors

Many key inputs to the development process are determined by the market and are fully or partially out of a local government's control

1. Target Return

Most developers do not have the $10s to $100s of millions necessary to fund a typical real estate project. Outside investors — such as a pension fund or Real Estate Investment Trust — provide most of the money for real estate projects and need a certain return, given the risks involved. Real estate projects are considered risky, and if the annual rate of return drops to levels close to that of less risky investments — such as Coca-Cola stock or "risk-free" US Treasury bonds — investors will choose to put their money in these less risky alternatives instead of funding this project.

Investors judge risk in a city based on certainty — if a number of past projects have been denied or faced significant delays, investors may require a higher return. Because return is calculated annually, and compounds over time, target return can make a large difference.

2. Landowner Willingness to Sell

To calculate whether a project will be profitable, developers often calculate all their costs the amount they would need to earn to pay back their investors and earn a decent profit, excluding the price of land. What is left over — profits of the project minus costs — is what the developer can afford to pay for land.

Property owners value their property differently, using the value of the existing use — the revenues generated by people paying to park on an empty lot, for example — to calculate how much it might currently be worth. To induce property owners to sell, real estate developers must often pay more than the property owner's perceived value. Property owners vary in the value of their current use and financial situation, so in any given area this value will vary. In a strong market, rising rents will increase the value of the existing use, meaning a developer must pay more for the land.

That's why in this model, as the market land cost increases, but the amount the developer is able to pay — Land Cost Paid / SF of land, to your right — remains the same, the property is less likely to be built. As the difference between what the developer is willing to pay and what the landowner sees in the market becomes bigger, the landowner becomes less likely to sell.

3. Local Rents and Costs

Local rents and sales prices, along with construction costs, make the largest difference in developer models. Small increases in rents can make a project very profitable, while small decreases can doom them, and vice versa for construction costs. And as market rents and sales prices rise, construction costs often rise as well due to skilled labor shortages, which means that rising rents do not always translate to increased feasibility. This is one of the reasons development is considered so risky and why investors, and developers, demand higher returns compared to other investment opportunities. Note this means that moving one of the levers here may also mean others may shift due to overall economic conditions, so users should be careful of the conclusions drawn from a large shift in any one of the indicators.

Local Government Factors

A number of important local government decisions can affect the odds that a project gets built, especially when these changes are large or the project in question has only a moderate probability of construction.

4. Fees or Affordable Housing Requirements

Developers may choose to build affordable housing on site — as part of the project — or pay a fee. Compared to local rents and investor risk adjusted return, fees and on-site affordable housing requirements are not as important to a development model. However, large changes to the requirements, and the affordability levels required, can have a significant impact on development in the short term. For example, requiring 20% affordable housing for lower income people at 50% of AMI or a $40,000 fee makes a much larger difference than requiring 10% of housing be affordable to higher-income at 120% of AMI or a $10,000 fee.

How a government implements the fees also matters. Developers negotiate an exclusive purchasing option with a landowner when first putting together plans for a site, which can happen a year or more before the development begins construction. If a policy does not grandfather projects during this time period, the developer will have to re-negotiate a lower price for the land in order to earn a profit. If the landowner refuses to sell at a lower price, the project may not be built. In addition, investors and developers may percieve the policy as unfriendly to development and require higher returns in the city in the future.

5. Local Planning Decisions

Though not in the same realm as local rents and investor risk-adjusted return, local Planning Department decisions on building height, parking, and permitting time can make a significant difference to a project's feasilibity. In this case, height refers to the height chosen by the developer, not a maximum height limit set by the Planning Department, in order to show the financial tradeoffs at each level of building height instead of just the most profitable height. Buildings over 6 stories often cost significantly more to construct because building codes require steel frame instead of the much cheaper wood, and life safety systems become more complicated. Parking can be very expensive — even with new technology, a single space can cost $25,000 to $75,000. Requiring less parking not only saves money but can allow for more space to build housing units, and for developers to earn money.

In the permit process, time is money. Though only a small chunk of the investor's money is spent up front, the investor still requires an annual return on that sum. The longer the permit process takes, the more profitable the project must be when it's finally constructed. Shorter permit times — such as shortening building permit wait times or plan checks — can mean significant savings.

6. Additional Planning Approvals

Requiring that developers obtain Conditional Use (CU) or Planned Unit Development (PUD) approval can be just as important to developers as local rents or risk adjusted return. Both are zoning designations created by cities that require a developer to get approvals from the local Planning Commission and possibly City Council in order to go through with development. As such, the general public gets to weigh in on the proceedings, and because the issue is now discretionary, projects must either be certified as having no environmental impact or go through the process of creating an Environmental Impact Report.

Issues brought up in these public meetings may be as small as requiring proper signs or security to appease neighborhood concerns, or they may be as large as providing additional affordable housing units or mitigating large environmental impacts, such as traffic issues. In the case of small requests, CUs or PUDs can be relatively harmless, if they are resolved quickly. In the case of larger issues, developers may end up spending over $1 million and 12 to 24 months on an environmental impact report, and additional time in court if sued by neighborhood groups, not to mention the additional costs of required mitigation.

As a result, investors and developers often require higher returns for properties in these zones, depending on the perceived risk and additional costs involved.

Other Factors to Consider

Many other factors affect whether a project gets built. For example, different assumptions for future rent or construction cost increases can radically change a developer's expected profit. Those willing to take a bigger gamble on larger rent increases could end up big winners — or out of cash.

Sources

This model draws inspiration from the very effective New York Times Buy Rent Calculator put together by Mike Bostock, Shan Carter, and Archie Tse. Key assumptions are drawn from numerous discussions with developers, contractors, and architects, as well as recent real estate feasibility studies done in San Francisco, Oakland, and El Cerrito by AECOM (here and here), Seifel Consulting (here and here), and Strategic Economics (here). Some basic assumptions for affordable condo purchase price calculations are drawn from the San Francisco Mayor's Office of Housing Inclusionary Housing Program (here). Structure and basic assumptions are also drawn from various real estate classes at UC Berkeley — taught by Bill Falik, Dennis Williams, and Carol Galante.

Methodology

The model calculates developer return using the Internal Rate of Return (IRR) metric. Model assumptions and calculations are based on a pro-forma analysis. For more detailed methodological information, contact ternerhousing@berkeley.edu.

The basic assumptions are as follows:

The condo or apartment project is sold at stabilization after all construction is complete at a rate of 20 units per month. Stabilization indicates the time when 90% of the building is occupied for 90 days for rental units.

Hard costs represent costs associated with actually constructing the building — digging the foundation, putting up the walls, installing the plumbing, etc. — while soft costs are all other costs — legal fees, some architectural fees, time spent meeting with city staff, and other costs.

The bank loan is 65% of total hard and soft construction costs and bank loan fees are 0.75% of the loan amount. Costs to build out the retail space are $15 per square foot (psf) less expensive than low-rise construction costs. Tenant improvements and leasing costs for the space are $100 psf. Basic podium parking — where part or all of the bottom floor is a concrete-walled parking lot — costs $120 psf. Low rise construction cost does not include retail tenant improvements, parking, or landscaping costs, and indicates the cost of building up to 5-story structures.

High rise construction costs are calculated internally by the model as a percent of low rise costs, rising to 103% of low rise costs at 6 stories, 125% at 7 and 8 stories, 130% of low rise costs at 9 stories, and 1 percentage point higher per story afterward.

Sales expenses are 3.5% of revenue for rental properties and 5.5% of revenue for condos. There is a premium for condo finishes of 5% of residential hard costs. All units are the same size in square feet. All units receive the same rent per square foot. There is a 10% contingency added to hard construction costs.

Each parking space requires 350 square feet, and stackers — allowing 2 cars to fit, stacked one on top of the other, in a single space — are used for an extra $15,000 per space for all parking in buildings 3 stories and higher. With stackers, parking costs come out to around $28,500 per space in a podium setup, compared to about $42,000 per space otherwise, a difference of over 30% from traditional parking costs. Construction takes 18 months for smaller projects 3 stories and higher, 6 months for 1 story projects, 12 months for 2 story projects, and 21 months for projects 7 stories and higher, adding one additional month time for every 2 additional stories for larger projects beginning at 13 stories.

The amount of lot square footage used as net rentable space is 73% for the most inefficient projects — high rises — and near 100% for one to two story buildings. Landscaping costs $10 psf. Above 50 units, projects gain efficiency at a slow linear rate up to a 5% discount on hard and soft costs for 250 units and above.

A number of these variables change with building height. Per square foot costs rise with building height according to general cost estimates from the San Francisco Transportation Sustainability Fee report and interviews with developers. Lot square footage used as net rentable space is 100% for 1 story projects and 95% for two story projects, dropping to 78% for 3 story buildings. The figure improves steadily from 5 to 7 stories, and then declines again after 9 stories in 1% increments. Parking costs per square feet are significantly cheaper for 1 to 2 story structures than more complex low-rise podium parking. Parking costs rise after 6 stories and again after 11 stories to reflect the costs of underground parking.

Residential vacancy is 5%, while retail vacancy is 10%. Soft costs, not including interest carry or local impact fees, are 20% of hard costs. The capitalization rate (cap rate) — termed Year 1 Return for Similar Properties above — for retail is 1.5 percentage points above the residential cap rate, as indicated in a recent CBRE report. Two parking spaces are required per 1,000 sf of retail space.

10% of hard and soft construction costs are spent during predevelopment. Predevelopment indicates the time before construction occurs, usually spent drafting plans, doing soil studies, getting city approvals, and other activities. 10% of the cost of the land is paid annually in option consideration.

Basic permitting processes includes time to get a planner assigned, design review, and time required to receive a building permit application. The EIR or negative declaration processes are assumed to be concurrent with basic permitting processes, while conditional use time is assumed to be above and beyond basic permitting time. We assume developers take about 3 months time before submitting anything to the planning department.

Projects with a negative declaration are assumed to take 4 months longer than projects with no environmental review, while projects with a full EIR take 12 to 24 months longer, depending on project size. EIR production and litigation costs are assumed to start at $300,000 and rise based on project size by about $1,500 per unit. Projects do not have to complete an EIR or negative declaration if they are not in a conditional use area, are not classified as a "large project" — here defined using the city of Oakland's definition of a lot size of 60,000 square feet or more — or have 6 units or fewer.

Affordable units are assumed to be the same size and quality as market rate units. AMI levels are based on 2016 Department of Housing and Urban Development income limits for a family of 4 in the Oakland-Fremont CA Metro Area. The density bonus is modeled after Oakland's ordinance — 10% minimum required at Low Income, with a 20% bonus, increasing by 1.5% to a 35% bonus at 20% affordable units. The model uses only the low-income scale, no matter what affordability level is selected. The model assumes there is no external subsidy for affordable units, though in practice developers might pursue a subsidy if more than 20% of units were required to be affordable at 50% of AMI or below. At this threshold, the project would qualify for non-competitive Federal Low Income Housing Tax Credit Subsidies.

Projects in liquefaction zones cost 4% more and projects in landslide zones cost 1.5% more, based on conversations with a structural engineer.

Condo fees are $500/month, and property taxes are 1.2%. Mortgage rates are 1 percentage point below the "Bank Loan Interest Rate" for construction loans. Affordable condo purchases pay 10% down. Renters and owners pay 30% of their income toward rent or mortgage payments. Operating costs are 30% of residential rents and 10% of retail rents.

The land seller in this model is assumed to be a motivated land seller. Land sellers are assumed to have different prices at which they are willing to sell (for those mathematically inclined, it is modeled here using a poisson distribution with lambda value 4 at increments of 0.05 from 0 to 10). The distribution, shown below, most closely matches the distribution of land seller willingness based on conversations with area developers. This assumes that about 79% of motivated land sellers agree to sell at the market price, allowing for a few holdouts. The percent of market land value is on the x-axis and probability of sale is on the y-axis:

This model works best at a scale of approximately 50 to 250 units, due to a number of unique factors in very small and very large projects. At a scale below 50 units, mid-size projects lose efficiency from a property management perspective, as managers have to split time between multiple properties. Costs may vary widely for smaller projects as up-front costs easily absorbed into larger projects become a larger share of the project budget. Single family homes and townhouses at a small scale have different rear yard requirements, ways of approaching parking, and requirements for building systems.

Project efficiency is assumed to increase by up to 10% of costs as projects rise from 50 to 250 units. Most defaults are set based on local data for a typical mid-sized lot in north Oakland, though some — such as target return and bank loan interest rate — are set based on conversations with developers. Local rents and sales represent the 85th percentile of north Oakland rents to approximate new construction.

This model does not include an ADU policy as a government lever because the factors that induce individual owners to build ADUs are too different from the factors facing larger building developers to include in the model.

© 2016 Terner Center for Housing Innovation at UC Berkeley. All Rights Reserved. No part of this website or computer code (javascript, HTML, CSS, etc.) may be reproduced, stored, modified, transmitted, or otherwise used without the prior written permission of the Terner Center for Housing Innovation at UC Berkeley.