Can Inclusionary Zoning Be an Effective and Efficient Housing Policy? Evidence from Los Angeles and Orange Counties

ABSTRACT:

Inclusionary zoning—requiring and encouraging developers to build some affordable housing in market-rate projects—is a growing but deeply contested practice. We evaluate the experience of inclusionary zoning programs in Los Angeles and Orange Counties, including their structure and elements, effectiveness in delivering affordable housing, and effect on housing markets and supply, to address the debate. We find that the programs vary but are not heavily demanding and include cost offsets. Low in-lieu fees, however, can be the weak link. Many of the mandatory programs are effective, if effectiveness is measured by comparing the affordable housing productivity of inclusionary zoning with other affordable housing programs. We found no statistically significant evidence of inclusionary zoning’s adverse effect on housing supply in cities with inclusionary mandates. We conclude that critics underestimate the affordable housing productivity of inclusionary zoning, and overestimate its adverse effects on housing supply. Nonetheless, inclusionary zoning is no panacea and needs to be part of a comprehensive housing strategy.

Notes

1 Inclusionary zoning and inclusionary housing are often used interchangeably in the literature and in practice. We, however, draw a distinction between the two strategies (also see Mallach, 1984). We consider inclusionary zoning as primarily a local government response that typically mandates, but sometimes just encourages, private developers to include some below market-rate housing in their projects. In contrast, we consider inclusionary housing as primarily a state government response aimed at preventing local governments from discouraging affordable housing developments in their jurisdictions. Thus, according to our definition, key examples of inclusionary housing include Massachusetts’ antisnob zoning law—Chapter 40B of Massachusetts General Law (Cowan, 2006), and New Jersey’s Fair Housing Act (Schwartz, 2006). There are, nonetheless, overlaps between these two approaches. For example, in California, the state government approved a Density Bonus Law. Accordingly, local governments are obliged to provide private developers a density bonus if they include affordable housing in their projects (Padilla, 1995).

2 In addition to the 17 cities (nine are in Los Angeles County and eight in Orange County), unincorporated Orange County also used to have a mandatory inclusionary zoning requirement that was introduced in 1979. It was replaced by a voluntary program in 1983.

3 In addition to the inclusionary zoning programs, many cities have active Community Redevelopment Agencies. As in the City of Los Angeles, these redevelopment agencies have independent affordable housing programs that include inclusionary requirements. This can also complicate the task of counting housing produced through a city’s inclusionary program, and might help explain some of the differences in the data collected by various researchers. For example, for cities like Laguna Beach and Monrovia, our research did not reveal much affordable housing production through the inclusionary programs but secondary sources list substantial inclusionary units in these cities. Our data indicate, and we list, four affordable units produced in Laguna Beach (and an unspecified amount of in-lieu fee collections) but Calavita and Grimes (1998, p. 159) list 310 units by 1998 and Benjamin Powell and Edward Stringham (2004b, p. 4) list 139 units. Inadequacies in city records and short affordability terms, as we discussed above, might also help to explain the discrepancies in the data.

4 Lake Forest’s 2000–2005 Housing Element specifies, but does not detail, a policy of “encouraging” the incorporation of a minimum of 15% affordable units within residential developments to help meet the city’s goal of having adequate housing to meet existing and future needs (see 2000–2005 Housing Element of Lake Forest General Plan, page H-4, adopted December 19, 2000). The city, however, has yet to institute the expected ordinance or further specify incentives for encouraging the set-aside. The Voluntary Incentive Program (VIP) established by Long Beach in 1991 is a three-tiered program. Its first tier mimics the State Density Bonus Law that was in place at the time of VIP’s adoption, and grants a 25% density bonus to projects with 20% or more of the total units reserved for low- and moderate-income households. The other two tiers offer a density bonus of 100% to projects that set aside all units for senior citizens and the disabled, or a 200% density bonus to projects that restrict all units for low-income senior citizens and the disabled. Affordable housing units created through the VIP must remain affordable for 30 years. The City of Monrovia adopted an Affordable Housing Owner-Occupied Incentive Program (AHOIP) in 1992. It also mirrors the State Density Bonus Law in place at the time of adoption, but only for ownership-based projects, and offers incentives such as the permitting of attached units, reduction in off-street parking, unit size reductions, less required recreation space, increase in floor area ratios, and modified setback standards. Nearly identical to the state program, AHOIP differs in its explicit listing of the available incentives, and its exclusive focus on ownership housing. Monrovia’s 2000–2005 Housing Element, however, makes reference to the city’s intent to extend the same incentives to affordable rental housing. Like Long Beach, all affordable units must remain restricted for 30 years.

5 In all jurisdictions, the negotiation of incentives generally requires approval of the planning commission. Although not every ordinance or housing element adopted by cities with mandatory inclusionary zoning lists the array of incentives available as Brea does, developers in all jurisdictions are entitled to request as many incentives as desired. Furthermore, the State’s new Density Bonus Law—SB1818—requires all jurisdictions to offer an appropriate density bonus, and one to three regulatory incentives (concessions) to facilitate the inclusion of affordable housing units, provided that the developer sets aside at least 5% of units for very low-income households, or 10% of units are set aside for low-income households. This minimum set-aside condition covers the mandatory requirements of all cities discussed in this analysis, except for San Clemente. San Clemente mandates only a 4% set-aside for very low-income households.

6 A few cities, including Pasadena and Santa Monica, have recently employed professional consultants to collect and analyze data to establish or adjust their fee schedules. For example, Keyser Martson Associates performed a financial analysis in October 2005 to update the fee structure for the City of Pasadena. Similarly, Hamilton, Rabinovitz & Alschuler, Inc. (HR&A) performed the nexus study for the City of Santa Monica in July 2005. HR&A focused on the demand for goods and services created by upper-income households purchasing or renting new market-rate units in the city. According to their analysis, delivery of these goods and services, in both the public and private sectors, to the upper-income households requires the employment of workers at all pay scales, including lower-income individuals that require housing at affordable prices. Thus, HR&A made the connection between the construction of new market-rate residential developments, and the need for new housing affordable to lower-income workers and their families. Based on this premise, HR&A performed an analysis to determine estimates of upper-income household spending, lower-income employment effects from that spending, the number of lower-income households associated with those employment impacts, and finally the appropriate affordable housing fee to offset the housing demand created by the upper-income households’ expenditures.

7 Construction cost estimates, according to developers we interviewed, vary between $150/square foot to $200/square foot. We assume the cost to be $175/square foot. For a supposed apartment of 1,200 square feet, we get a construction cost of $210,000. This estimate does not include marginal land costs.

8 Similarly, the Los Angeles Council members in their proposal for inclusionary zoning noted that “From 1980 to 2001, approximately 190,000 units were built in Los Angeles. If the City had a 15 percent set-aside requirement, through that time, 28,500 units of affordable housing would have been constructed” (Reyes & Garcetti, 2004).

9 A more recent report updated Brown’s numbers through July 2003 and found that the total production for the region exceeds 15,000 units (Fox & Rose, 2003).

10 Similarly, in Pasadena, the city has spent $2.3 million of in-lieu fee money to provide gap-financing for the development of 128 affordable units in two projects. The first is the “Trademark Project” consisting of eight units with a $1.3 million loan. The second is the “Heritage Square” project where the land to house 120 units was purchased for $1 million. The city was planning to release a Request for Proposals for developing Heritage Square, and planned to spend more of the in-lieu fees on the project’s development. Another potential problem in our analysis is that the listed totals of the in-lieu fee amounts might include other sources, such as fees collected from commercial developments through linkage fees. In Calabasas, for example, the collected in-lieu fees consist of residential payments (approximately 80%) and commercial impact fees (around 20%).

11 Most observers attribute the success of Irvine’s pre-2003 voluntary program to unique circumstances, including the Irvine Company’s relative monopoly on land development in the city and the threat of lawsuits (Calavita and Grimes, 1998).

12 The in-lieu fee in Rancho Palos Verdes was amended in late 2005 to $201,562 per affordable unit required (plus a 10% administrative fee), and in Calabasas it was increased in 2006 to $19–25 per square foot for market-rate units (see Table 3 for more details). Therefore, there is likely to be a substantial change in the productivity of their inclusionary programs in the following years.

13 Although our data set contains annual data for most cities, they are not a true panel data since we do not have data for all the years for each of our cities. Not all cities started inclusionary zoning in the same year. Additionally, we also had to substitute state level unemployment data for Orange County from 1980 to 1989 because county-level data were unavailable.

14 We used permit data from the Construction Industry Research Board for cities in Los Angeles and Orange counties. We have permit data from 1980 (which, for our data set, is also the first year when a city exercised inclusionary zoning policies) to 2005 for all cities, except four: Agoura Hills (data available from 1983); Calabasas (from 1991); Lake Forest (from 1992); and West Hollywood (from 1984). These cities were incorporated after 1980 and the permit data are available from their incorporation dates.

15 The variable COMEDU includes data from all cities for all years. A concern could be that including data from cities after they had instituted inclusionary zoning might distort their effect on the dependent variable. To avoid this possibility, we also created another variable for annual county median housing production in which median figures from years when cities had inclusionary zoning in place were excluded. Our t-tests indicated that there was no significant difference between these two sets of median figures.

16 For this, we use county employment data. We, however, could not find data for Orange County’s unemployment rates for the period 1980–1989, and use the state data for these years.

17 We also tested the effects of the independent variables separately on annual total single-family unit permits (SFU) and annual total multifamily unit permits (MFU) as dependent variables to test if the inclusionary requirements affect single-family or multifamily housing markets differently. But the outcomes of our models with SFU and MFU as dependent variables were similar to the results with TOTU as the dependent variable, and we present here and discuss only the latter.

18 We chose this variable (PCOTOT) over other potential local housing market variables, such as population or density, because it is one of the few proxies that provides annual figures.

19 The difference could be because the City of Los Angeles, which does not implement inclusionary zoning, greatly outweighs other cities in the county in terms of the number of housing units produced annually.

20 Our variable for the strength of the local housing market (PCOTOT) is potentially problematic. By defining it as a city’s annual total housing permits as a proportion of the total number of units produced in the county, we are able to construct an annual variable, but the dependent variable (TOTU) is also a part of this independent variable. This, however, does not necessarily mean that the two variables (PCOTOT and TOTU) are measuring the same thing, since a city’s housing production might change at a different rate from the county.

21 We recommend additional research that acknowledges the likelihood of interactions between local and regional elements and uses interaction variables to explore the market effects of inclusionary zoning policies.

22 In a similar way, the literature indicates that fast-track permits are valued by developers and can be attractive cost offsets. But if inclusionary zoning is mandatory, and all housing development projects qualify for expedited permitting, a city government needs to ensure that it commits the administrative resources for fast-track permits.