Building the next increment, by the numbers

Seth Zeren 

In the Strong Towns community we emphasize the importance of building to the next increment of intensity, allowing places to organically thicken up over time, flexibly responding to the current needs of inhabitants. I generally agree with this approach and try to put it into practice in my work as a neighborhood real estate developer. 

But there are some challenges we need to confront in order to consider how we might reshape our real estate markets to allow this sort of next increment of development model to be economically viable.

There are several reasons to like this model. First and foremost, it’s a strategy of many small bets rather than one big bet: meaning that if some things don’t work out as expected, the damage (public and private) is limited, and we can flexibly respond to try the next thing. For the same reason, these many small bets can be done by many small actors with fewer individual resources. We call this a city shaped by many hands. Small apartment buildings, duplexes, triplexes, and the like can be owned by mom-and-pop (small) landlords, owner occupants, and local bank debt. Massive 200-unit apartment buildings can only be owned by large institutional landlords. 

When land is rezoned for a big jump in density (light industrial to 4-story multifamily, e.g.) there is also a large increase in the land prices due to the new development potential. If land is priced for the largest project that can be built, then it will be too expensive (per new built square foot) for smaller, more incremental projects. 

If there’s not enough demand or development capacity to build more than one larger building every few years the result of a large upzoning may be less development than in the case of a small upzoning. This is because land prices increase and the market adapts to the new opportunity and constraints. So non-incremental development may tend to encourage speculation vs. value-added development. It may also tend to concentrate large investments into a small percentage of the available land. One parcel gets a new shiny apartment building, while nine other parcels around it continue to decay while waiting their turn. 

So this all sounds great, but there is a serious problem when it comes to how it would actually be implemented on actual parcels. Exploring one example should be illustrative of the pitfalls and the opportunities in pursuing next increment development. 

This two-family house at 1252 Westminster Street in my neighborhood in Providence, RI recently sold for $460,000 (if you allocate about 20% to land value, that’s around $153/SF for the house). My guess is that an owner occupant bought it to live in during our current crazy housing market and I doubt they have any redevelopment ambitions. But let’s imagine that we had bought the property with an intent to redevelop it. The property is located on a busy main street in a desirable, “gentrifying” neighborhood with solid amenities. Under current zoning, we could build up to three units, so we could add one more – sounds like a good incremental step!

Option 1: It’s not a great looking building, especially for a mainstreet–let’s knock it down and build three new units. For the sake of argument, let’s say we’re building three 1,500 square-foot flats (condos generally have a higher sale price per square foot than rental apartments).

The resulting building is three times as expensive per square foot than the existing structure! I think it’s just barely possible that our market would support something like these condos if they were really nice. 

On balance, a move like that would still be helpful to the city in terms of building more, nicer units to take pressure off other parts of the housing stock. Building it would improve unit quality, on average, creating more family-friendly housing. But it’s also not hard to see why plenty of people would look at that and see that it’s obviously bad for affordability. Of course, in a really strong market with more restrictive zoning, you would buy a house like that, tear it down and rebuild a McMansion at around the same price per square foot with only one-third of the density.

Option 2: The existing building has a lot of value that we’re throwing away by tearing it down. What if we just heavily renovated it and built a 2,000 square-foot addition (add a floor, build off the back, something). We’re still going to upgrade the existing building quite a lot to bring it up to modern standards, but we’re not throwing away the structure or foundation. 

This is a bit better, a 12-percent savings vs. the tear-down scenario. The savings increase to about 20 percent if the rehab cost goes down to $100 per SF, which is possible if you’re really frugal. On the other hand, I doubt very much that the resulting units would be able to get the minimum required sales prices! Much better to sit on your duplex and rent it while prices appreciate around you!

Option 3: What if we could get more out of the purchase price? Leaving aside the current zoning requirements, let’s try a small three-story apartment building with twelve smaller 1,000 square-foot apartments. This is more like two steps of intensity than one, but let’s consider it. The additional density means this building probably requires significant utility upgrades and because it’s a small lot, it has no off-street parking and little to no yard. How does this project fare?

This project hasn’t reduced our cost per square foot over the second option, though the smaller 1,000 square-foot units would be less expensive than the larger units in the first two options. You couldn’t do this project under current zoning for a host of reasons, but if you could design it and get it permitted, would it work as a for-sale project? Maybe… if they were really nice. But you couldn’t get financing to build it unless it also worked as a rental building. 

Rental buildings are valued based on their Net Operating Income (NOI = Revenue minus Expenses, before debt). The ratio between NOI and sales prices is called the Cap Rate. A cap rate of 6.5 percent would be pretty typical in our market for a building like this. Then NOI would have to be around $21.67 per SF per year. If operating expenses are 35 percent of gross revenue (at typical figure), then the gross rents would need to be about $33.34 per square-foot per year, which translates to a monthly rent of about $2,775 per month for a 1,000 square foot apartment. That’s pretty high for Providence even on the upper end of quality. Maybe you could do it, but it would be quite risky.

Takeaway Lessons

Existing buildings, even pretty junky buildings, have significant value. When you lose some or all of this value through redevelopment to the next increment of intensity, it substantially increases the cost of development relative to doing nothing. It may make incremental redevelopment infeasible. A high-price, low-quality existing land use (e.g. a Dunkin Donuts with a drive-through) is very hard to redevelop even with a big building. It is basically impossible with a smaller, more incremental step: there’s too much “value” that would be destroyed doing something different. 

This is a major reason gentrification happens where it does. You need a big gap between current prices and potential prices in order for any development to happen. It’s in these rent-gap seams that development is feasible. Since that’s where development can happen, gentrification is associated with development, but in a sense it’s backwards. Development follows the gentrification opportunity; even as it sometimes helps generate the snowball of demand which propels the process. Current prices often have to fall very low to get to the point where redevelopment makes sense. 

Demolishing existing buildings throws away a lot of potential value. Structures that are more durable and flexible are very valuable for your community because they can adapt to changing demand and uses without being torn down. Buildings which can learn with time and be affordably added on to or renovated are very valuable. Keeping them is a long-term strategy to be encouraged in your community. 

The other major variable here is the expense of new construction. It’s typical to pay more than $200 per square foot in our market. At this cost per square foot, only expensive, high rent buildings can get built and the gap between current intensity and the next step has to be larger. If we could lower the cost of construction, development to the next increment would be more feasible. I plan to explore strategies we could use to lower construction costs and increase our capacity for building things in a future article. 

I come away from this analysis with a lot of questions. Strong Towns has often shared the montage of photos from the development of Brainard, Minnesota, over the decades. How did their business models work? How were properties priced? How were they financed? How big of a step in value was needed to justify each additional increment? 

We have a lot to learn from the past – not just about design, but about the business models and practices that allowed for development to the next increment of intensity.