Zoning Constraints & Housing Types

We all know zoning restricts housing supply in cities. However, the type of housing produced will be different for different kinds of zoning regulations. In this post, we’ll explore the impact of three common kinds of zoning regulations: density controls (number of units), height and setback requirements, and floor area ratio (FAR maximums). As we’ll see, while variety of housing is often a stated goal of planning, zoning regulations and market conditions often work to the contrary. Height and setbacks work in the same way as FAR, with one always being more constraining than the other for a given lot.

Method of Analysis

To simplify things, we’ll look at the impact of these three types of regulations on a 50’x150’ (7,500 square foot) lot, which can be found all over LA and Glendale. For LA, we’ll consider the R1, RD3, RD2, RD1.5, R3, R4, and R5 zones as defined by the city of LA. For Glendale, we’ll consider the R1, R3050, R2250, R1650, R1250, and SFMU zones (which roughly correspond to R1, RD3, RD2, RD1.5, RD1.5, and R4). We will look at the number of units and size of building possible on a 50’x150’ lot in each zone, and see the impact on the type of housing produced.

In general, we will see that the lower density zones are constrained by permitted density, which tends to result in the production of only large, expensive housing units. High density zones are constrained by height & setbacks or FAR, which tends to result in the production of only one bedroom (1BR) and two bedroom (2BR) units, leading to the charge that apartment developers don’t build for families.

Los Angeles

The table below summarizes the maximum permitted density, setbacks, and FAR in common residential zones in LA, assuming height district 1L, except for R5 where we assume height district 2, for reasons explained below.

Again, assuming a 50’x150’ lot, the maximum number of units, maximum floor area, and average floor area per unit are as follows. Assumed efficiency means the percentage of building floor area that’s actually usable for apartments. For single-family structures, it can be assumed to be 1.00. For apartments we assume 0.80 for a low-rise apartment in the R3 zone, and 0.70 for mid-rise apartments in the R4 and R5 zones. Efficiency for apartments is less than 1.00 because of space lost to hallways, elevators, common areas, trash rooms, and so on.

LA’s FAR is very generous for low density zones, so height & setbacks rather than FAR end up constraining maximum floor area for all zones except R4. If we had used height district 1L for R5, it would also be constrained by FAR instead of height & setbacks, and would only have an average unit size of 425 SF.

As a practical matter, in the R1, RD3, and RD2 zones, actual building size will be constrained by market conditions. There just isn’t that much demand for houses over about 3,500-4,000 SF. These zones are purely constrained by density, meaning that developers will max out the number of units possible and build the largest units they think the market will accept. Purple City once ran the numbers to show you why developers won’t put small houses on big lots.

The RD1.5 and R3 zones are more or less equally constrained by density and building height & setbacks. For R3, density has increased to the point that average unit sizes have been driven down to about 2,000 SF for a small lot subdivision of free-standing houses and about 1,600 SF for apartments, housing unit sizes that are in high demand. This is probably one reason the R3 zone is popular with small lot developers; the combination of permitted density and floor area doesn’t force the units to be smaller than people want, nor does it force much of the lot to remain as open space.

The R4 and R5 zones are constrained by floor area, whether in the form of maximum FAR or height & setback requirements. If the developer maxes out the number of units, they will only be able to get about 800-900 SF average unit size. This is why large apartment buildings in LA are almost all studios, 1BRs, and 2BRs. If you tried to make a decently-sized 3BR unit, say 1,400 SF, it would have to offset by two units of only 500 SF, or a reduction in total units.

Note that if a development is FAR constrained, parcel assembly doesn’t help with unit size at all, only with making it easier to configure parking ramps, elevators, and other common spaces. If a development is height & setback constrained, parcel assembly will help with unit size by eliminating setbacks between lots, until the point FAR constraints take over.

Glendale

The analysis is similar for Glendale, but maximum FAR in Glendale is much less, and setbacks and heights are more restrictive. The table below summarizes the maximum permitted density, setbacks, and FAR in common residential zones in Glendale. Setbacks are averages because Glendale has step back requirements for second and third floors.

Again, assuming a 50’x150’ lot, the maximum number of units, maximum floor area, and average floor area per unit are as follows. I’m assuming 0.90 efficiency for townhouses.

Except for R1250, the multi-family residential zones in Glendale are in the sweet spot for townhouses (1,500 SF to 2,000 SF). The R1250 zone would work for small townhouses or 2BR apartments.

For lots over 90’ wide, Glendale allows additional density and another story of height in the R2250, R1650, and R1250 zones. There’s also a mixed-use zone, SFMU, that requires 100’ wide lots. Therefore, the analysis is modified if you assemble two lots. The SFMU zone has maximum height of 60’/4 stories and density 87 units/acre when abutting another multi-family zone, and 75’/6 stories and 100 units/acre when not, so results are presented for both cases. In practice, it is very rare for an SFMU zone to not abut another multi-family zone. The given story heights for SFMU assume half of the first floor is retail space and while max FAR is not specified it can be inferred from story height multiplied by 0.9, since 10% of the lot must be landscaped.

Because density is increased but FAR is not, the average unit size is actually driven down, despite being allowed to make the building one story taller. Of the few multi-lot townhouse projects I’ve followed in Glendale, many of them have not maxed out the density in these situations, electing to build fewer, but larger units. A motivating decision here is probably Glendale’s requirements for 2 subterranean parking spaces per unit, so density may actually be maxed out based on the number of parking spaces you can build in one underground level.

The SFMU zone ends up with larger average unit size than LA’s R4 and R5 zones, and sure enough, you do see some 3BRs in new developments in downtown Glendale. (While not actually in the SFMU zone, most of these buildings are in zones that allow 90-100 units/acre and up to 6 stories by right, so they’re a reasonable proxy.)

Encouraging Housing Diversity

Certainly, cities could increase the diversity of housing production by liberalizing zoning. Increasing allowable density and FAR, and eliminating minimum unit sizes, would allow different developers to try more different kinds of projects. After all, it was more liberal zoning regimes that produced neighborhoods that have a wide variety of housing types, like South Glendale.

Failing that, there are some other policies that might help. The primary concerns seem to be that apartment builders do not build enough family-sized apartments, while townhouse and small-lot builders do not build enough small homes. Some possibilities:

• Give apartment developers free FAR for every bedroom beyond the second, for a certain percentage of units. Height and setbacks would have to be generous enough to make the extra FAR usable.
• Add a density bonus for providing 3BR or 4BR apartments; for example, allow 0.20 additional units for every 3BR and 0.30 additional units for every 4BR, up to a maximum. FAR, height, and setbacks would have to be generous enough to make the extra FAR usable.
• For townhouses and small-lot subdivisions, rezone outlying R1 areas as RD1.5 or R1250. Land in outlying areas is cheaper, reducing the need to max out FAR.
• Add a density bonus for building small townhouses or small lots; for example, in the RD1.5 zone, allow 1000 SF lot area per unit up to certain percentage of units if they are smaller units.

 

Advertisement

LACMTA Valley Bus Ridership – September 2016

Here’s our fourth update on ridership on some of the main bus routes in the San Fernando Valley. As a reminder, for north-south corridors, we have San Fernando, Van Nuys, Sepulveda, and Reseda; for east-west, Ventura, Sherman, Roscoe, and Nordhoff.

For more detail on the sausage-making involved in converting routes that cover multiple corridors to a number for a single arterial road, see the first post.

Here’s the raw data. As always, highlighted cells represent top 10 ridership months since January 2009. All routes put up their best months in the 2009-2010 period; this may be due to the recession reducing car ownership.

Here are the 12-month rolling averages for weekdays.

Saturday and Sunday 12-month rolling averages largely reflect weekday trends, as shown below. The previously noted uptick in Reseda ridership on weekends has reversed.

As discussed previously, the configuration of rapid routes on Van Nuys was changed in late 2014. Route 761, a rapid that went from Van Nuys in the Valley through Sepulveda Pass to UCLA in Westwood, was eliminated. At the same time, Route 734, the Sepulveda rapid, was extended from its previous terminus in Sherman Oaks through Sepulveda Pass to Westwood. Rapid service on Van Nuys was replaced with Route 744, a U-shaped route on Van Nuys, Ventura, and Reseda. An express rapid service, Route 788, serving the northern part of Van Nuys and connecting to the Orange Line, then running express on the 405 to Westwood, was also created.

Here is the breakdown of weekday ridership on Van Nuys and Sepulveda by local and rapid on each corridor, and total local and total rapid on the two corridors combined.

Since a longer time has passed, we can now also start to look at the 12-month rolling averages.

The rapid route shuffle seems to have not had much impact on overall ridership trends. Weekday local ridership had already begun to trend down when the shuffle took place.

In contrast, it seems possible that weekend ridership has suffered. While Route 761 ran on weekends, Route 734 never has, and this was not changed when 761 was eliminated. Route 744 runs on weekends, but Route 788 does not; thus on weekends there is now no rapid service from the Valley to the Westside.

Again, we are speculating, but it appears that with the elimination of 761, riders who couldn’t cancel their trips and had no other option to get from the Valley to the Westside shifted to the Sepulveda local route, 234, producing a sudden jump in ridership. The increase in local ridership was smaller than the drop in rapid ridership, so overall ridership has trended down. However, the background trend has been a decline in ridership, so while possible, it is cannot be said with any certainty that the rapid route shuffle caused a decline.

Transit, Ride-Hailing, & Class-Mixing

As venture capital-backed ride-hailing services like Uber and Lyft continue to expand, there has been a lot of speculation on the impact of these services on transit. Will they replace transit services, as riders defect to faster car trips, or will they complement transit services, as riders use them for last mile connections? And, if riders who can afford to defect to ride-hailing services do so, will that lead to a vicious cycle of worsening transit, as decreasing ridership and political leverage cause further reductions in service?

On the first question, time will tell, but it seems like things could go either way. In congested cities, transit has considerable geometric advantages over cars, provided it has its own exclusive or semi-exclusive guideway. However, if transit does not have its own right-of-way or lanes, it offers little advantage over driving, and ride-hailing trips might replace transit trips. This could lead to a socially suboptimal Nash equilibrium, where everyone would be better off if some people took transit but no individual has the incentive to do so. (Ignore, for simplicity’s sake, the potential to introduce congestion charges, or the question if ride-hailing services will be able to scale and be profitable.)

In addition, many smaller cities in the US do not suffer from appreciable congestion, and in these places transit’s geometric advantages are less relevant. Again assuming they can be operated profitably, ride-hailing services might be able to capture some trips in these cities as well.

Does that spell disaster for transit services? I don’t think so. Voters in many US cities have shown their willingness to increase their own taxes to fund capital improvements to transit, even in cities with relatively low transit mode share like Los Angeles, Denver, and San Jose. While funding for operations and maintenance remains a major issue for many agencies, it doesn’t seem unreasonable to think that voters could be persuaded to fund O&M as well. (In LA, at least, some funds from voter-approved measures do go to operations.)

There is also concern that loss of ridership to ride-hailing services would reduce mixing of classes that occurs on transit but not in other transportation modes. Transit itself usually already has an informal hierarchy that separates classes, with commuter rail at the top, followed by rapid transit, and then local bus. (There’s even stratification within modes; I’ve had people tell me why the Ventura County Line is a better Metrolink line to ride than then Antelope Valley Line.) So ride-hailing services may reduce class mixing, though mixing and interaction are not the same thing. A person is probably more likely to talk to their taxi driver or ride-sharing companion than a random person on a transit vehicle.

However, even interaction does not compel understanding. It’s usually remarkably easy to get people to open up and talk about their lives if you want to listen. It’s even easier to just make small talk, or not talk at all. Meaningful interaction with different people only happens if we want it. Expecting a transportation technology to make it happen seems about as fruitful as expecting ride-hailing technology to solve our poor land-use policies.

LACMTA Bus Ridership Update – August 2016 Edition

Six months have passed, so it’s time for another LACMTA bus ridership update. As always, we start with the raw data. Highlighted cells represent the top 10 months for that route (since January 2009).

Since the Expo Line to Santa Monica opened during this time, I thought it might be good to look at the monthly data in addition to 12-month rolling averages. Here are the weekday, Saturday, and Sunday raw data graphs.

Here are the weekday, Saturday, and Sunday 12-month rolling averages.

It’s impossible to say what the impact of the Expo Line is without polling riders; however, there is not a large change in the trend for any line except Wilshire. There is a seasonal drop in Wilshire ridership data during the summer, but it looks larger than normal this year. Looking at the Wilshire split data between routes 18, 20, and 720, it looks like there was a drop of a few thousand riders in 720 ridership after the Expo Line extension opened. The Expo Line would be a shorter ride from downtown LA to Santa Monica than route 720. Again, we cannot say if this is what happened without actually asking riders.

There’s not much else new to say, so we’ll keep it short. Lines that have seen slight decreases continue to decrease; those that are steady seemed to keep holding. The Silver Line continues to grow slowly.

Here’s the percentage of trips on each arterial being served by the rapid route.

The share of riders served by the rapid routes continues to slowly rise on most corridors. This doesn’t necessarily mean increasing ridership on the rapid – it could be that both the rapid and local declined, but the rapid was more resilient.

That’s it for now; next up, Valley bus ridership.