How much infill can we build in California?

This is the big question everybody is trying to answer these days. Now, in typical fashion, UC Berkeley planning professor John Landis has taken a crack at answering the question in a comprehensive – and somewhat controversial – manner.

In a major study conducted for Business, Transportation and Housing Secretary Sunne McPeak, Landis has concluded that about 25% of California’s future housing needs – about 1 million units -- can be met with infill development. (A summary of the report can be found at

There are many caveats in Landis’s study, but that’s the bottom line: Statewide, 75% of all new housing will have to be built on greenfield sites. And that’s using a pretty generous definition of infill sites that includes, basically, all undeveloped land inside current city limits statewide.

Given all the talk about infill these days, 25% doesn’t sound like much. In fact, the total size of all the infill sites Landis identified statewide would fit inside the San Diego city limits. Can this really be the upper limit?

The answer is yes and no. Yes in the sense that Landis has identified most of the obvious infill sites in the state. But no in the sense that he calculated the potential number of units based on his definition of current economic feasibility. The economics of infill may change in the future – and so might allowable densities under zoning, which is something Landis did not take into account in the study.

The Landis method isn’t perfect, of course. The list undoubtedly includes thousands of parcels that local governments would not consider to have infill potential. The lot I live on is on the Landis list, even though the house was expanded from a ’20s bungalow to a large duplex in 2002. The building where our offices are located is also on the list, largely because it is an historic structure that has been owned by the same family for a long time, and, therefore, the building is worth much less than the land, according to our county assessor.

The Landis report is unlikely to prove useful for the other goal that Secretary McPeak had for it – as a resource for developers and local governments seeking to identify infill sites. Anybody can use the “infill parcel locator” that Landis developed, which can be found at But because it uses a standard statewide definition of infill and does not include zoning, it probably won’t screen out the political problems.

Nevertheless, the big picture that emerges from the Landis report is worth discussing in more detail – largely because Landis has definitely confirmed some of the most important assumptions about California’s infill potential and why developing infill projects is so different. Here are four issues worth contemplating – along with one big surprise.

1. Most of the infill potential is in Southern California.

Most infill experts have long suspected the great infill potential lies in Southern California – especially Los Angeles and Orange counties, with their vast store of postwar commercial strips. For the first time, Landis has confirmed this. According to his estimates, 70% of the infill potential is located in metropolitan Los Angeles (not including San Diego). This compares with only 20% in the Bay Area.

2. Most parcels are not vacant.

The homebuilding industry often seeks to define infill broadly to include a supposedly vast array of passed-over urban land that has never been built on – rather than just reusing land that has already been urbanized. But Landis shows that the vacant passed-over parcels account for less than 30% of all land with infill potential. And it’s a fair bet that most of the vacant land has already been picked over by large developers seeking to apply their greenfield models to infill sites.

3. The parcels are small – unless they are industrial.

Landis found about a half-million parcels of land in California with infill potential. But the average size of the half-million parcels is only about 15,000 square feet – meaning that all the parcels combined add up to 340 square miles, or about the size of the City of San Diego. Furthermore, the more infill is defined in the classic “reusing existing land” sense, the smaller the parcels become. The average infill site that already has development on it – what Landis calls a “refill site” – is only about 15,000 square feet. The average vacant infill site, by contrast, is more than an acre in size.

Interestingly, commercial properties with infill potential are not much bigger than residential properties – half an acre on average. By contrast, industrial infill sites average 1.3 acres in size, confirming a widespread belief that industrial properties will be the next big “play” for infill housing developers. Recent research here at Solimar Research Group (CP&DR’s parent company) has confirmed that industrial properties are undervalued as infill housing sites, and anecdotal evidence suggests that this is where infill developers are headed next.

4. It’s hard to make rental projects pencil – and maybe not advisable.

Landis also came up with another finding that squares with what almost any infill developer will tell you – rents are so low compared with real estate prices, and building costs are so high, that it does not make economic sense to do infill rental projects. There’s a pro and a con to this. The con is that, by any measure, rental housing is the most important housing need in the state, given the modest incomes and increasingly working-class nature of the economy. The pro is that economic infeasibility will save the state’s current renters from a lot of disruption. Landis found that about 30% of the state’s infill parcels already have rental apartments on them. Redeveloping those parcels would replace high-density housing with higher-density housing – affordability might be reduced – and it’s probable that California would see an endless number of knock-down-drag-out fights over evicting current tenants and replacing them with new tenants.

5. The surprise: The infill market is Latino families, not couples without kids – at least in L.A.

In addition to measuring infill potential, Landis also tried to examine where infill demand is coming from. He did so by identifying the “cohorts” that were moving into infill-type neighborhoods. In the L.A. area, most of these cohorts were young Latino families of modest incomes. This contrasted dramatically with San Diego and the Bay Area, where the cohorts are mostly affluent white folks without kids. This is no surprise to anybody who has talked to an infill developer in L.A. during the last five years, but it does suggest a bifurcation of the state’s infill market: Latinos in L.A., classic yuppies elsewhere.

Given the limitations of Landis’s database, it’s hard to know exactly what the state should do with the wealth of information his report has produced. At the end of the report, he makes a bunch of pretty standard policy recommendations – gather better data, streamline the California Environmental Quality Act – to overcome barriers to infill.

But maybe the state should be more bold. Landis acknowledges that his study is an examination of what “could” happen, not what “should” happen. This is partly a sop to the local governments, who were a bit scared by the idea that HCD was compiling a statewide database of infill sites. Maybe, however, the state ought to move from “could” to “should” by adopting Landis’s assessment as a target. What if the state passed a law declaring that, as a matter of policy, 25% of all housing in the state should occur in infill locations? Then all the other recommendations would have to fall into place.