---

Summary

The Legislature recognizes that the availability of housing is of vital statewide importance and that the State and local governments have a responsibility to facilitate the development of adequate housing. State law requires the Department of Housing and Community Development (HCD) to conduct assessments to determine the housing needs (needs assessments) throughout regions in the State. The needs assessments rely on projections of future population and households developed by the Department of Finance (Finance). HCD is required to consider certain factors identified in state law and then can adjust the needs assessments for any of the factors. For example, it makes an adjustment to achieve a healthy vacancy rate in the housing market and an adjustment to reduce the number of overcrowded households. Regions use the needs assessments to plan for additional housing to accommodate population growth and address future housing needs.

---

HCD’s Housing Needs Assessment Process Lacks Sufficient Reviews and Support

HCD does not have a formal review process for the data it uses to determine its needs assessments. As a result, the needs assessments for two of three regions we reviewed included errors. One data error reduced a region’s needs assessment by nearly 2,500 housing units. HCD also did not demonstrate that it adequately considered certain factors when creating the needs assessments of the three regions we reviewed. For one of those factors, the healthy vacancy rate, HCD did not perform a formal analysis to adequately support its assumptions. HCD’s insufficient oversight of its process and the lack of adequate documentation supporting the healthy vacancy rate risks eroding public confidence in HCD’s ability to address the State’s housing needs.

---

Finance Provides Reasonable Population Projections, but It Has Not Provided Sufficient Support for Its Household Formation Projections

Finance’s projections of the statewide future population are reasonably accurate, but it did not sufficiently support its projections of the number of future households. To calculate the household projections, Finance identifies rates at which it expects individuals in different age groups to form new households and applies those rates to its population projections. Although Finance worked with HCD to solicit some advice from experts when it established these rates, it did not conduct a formal study or receive clear recommendations to support them. As a result, Finance cannot ensure that it is providing the most appropriate information for HCD to include in its needs assessment process. Finance stated that it intends to reevaluate its assumptions related to household growth after it reviews 2020 Census data when those data become available later this year.

---

Introduction

Background

As part of the Legislature’s efforts to ensure that the State is planning for the construction of enough homes to meet its housing needs and that local governments are facilitating that development, state law requires the Department of Housing and Community Development (HCD) to conduct periodic housing needs assessments to determine existing and projected housing needs throughout California. HCD fulfills its responsibilities under state law by creating Regional Housing Needs Assessments (needs assessments). As Figure 1 shows, HCD provides the needs assessments to councils of governments, which we describe in the text box, across the State and directly to counties that are not in such a council. Figure 2 provides an overview of the councils of government in the State and also shows counties that are not part of a council. After a council of governments receives its needs assessment from HCD, it then must allocate the region’s housing needs to the cities and counties within its boundaries. For counties without a council of governments, HCD provides allocations to those counties as well as to the cities within them. ¹ Counties that receive their assessments and allocations directly from HCD represent just 3 percent of the State’s population. Cities and counties must then develop plans to accommodate the existing and projected housing need. HCD performs needs assessments every five to 11 years. HCD does not complete all assessments at the same time and does not always cover the same period, because it attempts to align the needs assessment process with other planning processes, such as regional transportation planning. The three needs assessments that we reviewed are those of the Santa Barbara County Association of Governments (Santa Barbara Association), the Sacramento Area Council of Governments (Sacramento Council), and Amador County.

Figure 1

HCD’s Housing Needs Assessments Inform County and City Housing Plans

Figure 2

Most California Counties Have a Council of Governments That Receives Needs Assessments From HCD

Needs Assessment Components

State law requires HCD to use population projections developed by the Department of Finance (Finance) when it completes the needs assessments. Finance factors into its projections multiple sources of information, including data from the U.S. Census Bureau (Census) and records of driver’s licenses, births and deaths, school enrollments, and tax filings. Finance provides state‑ and county‑level population projections to assist state, regional, and local planning, among other purposes. Finance also projects the number of future households, based on the population projections and the percentage of people in the population who are expected to form their own households in the future, which is known as the household formation rate.

Table 1 describes the factors that state law requires HCD to consider in its needs assessments, including vacancy rates. State law requires HCD to consider vacancy rates in existing housing and the vacancy rates for healthy housing markets when developing the needs assessments. A low supply of housing can result in low rental vacancy rates, which in turn can lead to housing price increases. Therefore, HCD adjusts its needs assessments so that housing markets can achieve a healthy vacancy rate. In some cases, that adjustment will add to the number of housing units HCD determines a region needs so that the region can obtain a healthy vacancy rate. State law specifies that the minimum vacancy rate for a healthy rental housing market is 5 percent, but the law does not define the healthy vacancy rate for owned housing.

Table 1

Factors HCD Must Consider in Its Assessments

FACTOR	DESCRIPTION
Anticipated Population Growth	Projection of future population growth in the region.
Household Formation Rate	The rate at which individuals form new households in the region.
Household Size	The number of people per household in the region.
Vacancy Rates	The percentage of homes available for rent or sale compared to the total number of housing units, less vacation and seasonal homes.
Overcrowding	The percentage of households that have more than one resident per room in a housing unit.
Replacement Needs	Replacement of housing units lost during the planning period, such as because of deterioration.
Cost-Burdened Households	The percentage of households that are paying more than 30 percent of their income on housing costs.
Units Lost to Emergencies	The loss of housing units during a state of emergency declared by the Governor, such as in wildfires, if the lost units have not yet been rebuilt or replaced.
Jobs/Housing Balance	The relationship between the number of jobs in a region and the number of housing units in that same region.
Other Characteristics	Other characteristics of the composition of the projected population.

State law also requires HCD to adjust its needs assessments to account for long‑term housing challenges, such as overcrowding, which occurs when a housing unit has more than one resident per room. The Legislature added this overcrowding factor to the needs assessment process in 2017. HCD must also consider cost‑burdened households, which are households that pay more than 30 percent of their income for housing costs. When it determines it is appropriate to do so, HCD includes in its assessments adjustments for cost burden and overcrowding. Among the sources HCD uses to determine these adjustments is data that state law requires councils of governments to provide. The councils provide data comparing the cost burden and overcrowding for their respective regions with that of other comparable regions in the United States. HCD then uses this information to calculate adjustments for each council of governments’ needs assessment. Table 2 shows a hypothetical example of how HCD incorporates adjustments for the various factors to determine the number of housing units in its needs assessments. Appendix A shows the three needs assessments that we reviewed.

Table 2

Housing Needs Assessments Contain Many Factors and Adjustments

	HYPOTHETICAL EXAMPLE OF HCD NEEDS ASSESSMENT CALCULATIONS
	FACTOR/SOURCE	PROJECTED CALCULATION
JUNE 2020–JUNE 2028 (8 YEARS)	8-year Population Projection (Finance)	1,500,000
	– Group Quarters Population (Finance)*	– 35,000
	Population Needing Housing (Finance)	1,465,000
	Household Formation Rate Adjustment (Finance)†: 36.6% average
	Projected Households (Finance)	540,000
	+ Vacancy Rate Adjustment (HCD): 2.2%	11,900
	+ Overcrowding Adjustment (HCD): 0.6%	3,200
	+ Replacement Needs Adjustment (HCD): 0.5%	2,700
	Units Lost to Emergencies (HCD)‡	—
	Jobs/Housing Balance (HCD)‡	—
	– Occupied Units (Finance)	– 480,500
	Subtotal	77,300
	+ Cost Burden Adjustment (HCD)§: 0.55%	3,100
	Total Needs Assessment	80,400 Housing Units

Finally, state law requires HCD to consider housing units that communities will need to plan to replace. Some housing units become uninhabitable during the future period covered by the assessments, such as housing lost due to damage, deterioration, and house or apartment building fires. State law requires HCD to review housing replacement needs, and HCD does so by obtaining from Finance the number of housing units a council of governments or county has lost over the past 10 years. HCD then determines the rate at which the region loses housing units and makes an adjustment in the needs assessment to replace those houses. In response to recent wildfires that have destroyed a significant number of houses, the Legislature added the requirement in 2018 that HCD must also consider any housing recently lost during a state of emergency that the Governor declared. Similar to the cost burden factor discussed above, state law requires councils of governments to provide data to HCD on housing lost during a state of emergency for consideration in the needs assessments.

Local Actions After HCD Completes a Needs Assessment

After HCD makes a final determination for a needs assessment, state law requires the council of governments to create housing needs allocations for the cities and counties within its region. The council, in consultation with HCD, must develop a proposed methodology for distributing the allocation. The council of governments must conduct a survey and ensure public participation when developing the methodology. The council of governments establishes a draft allocation and then may hear appeals of the allocation, if any are raised. It then must make the allocation final and adopt it.

State law requires local governments, such as cities and counties, to create plans to meet housing needs. Local governments must adopt a general plan, which is a blueprint for meeting the community’s long‑term vision for the future. Within the general plans, state law requires local governments to include a housing element, which contains an analysis of existing and projected housing needs in their communities. Cities and counties must state their goals, policies, and programs related to the development of housing, to accommodate projected housing needs allocated by their council of governments or HCD. The community, through the housing element, must attempt to meet these housing needs, such as by changing the zoning on specific parcels to allow residential development.

Needs Assessments Can Be Contentious but Are a Critical Component of Addressing Housing Challenges

Some stakeholders have criticized the needs assessment process and HCD’s needs assessments. For example, some homeowners and advocacy organizations believe that HCD’s needs assessments have produced higher numbers of housing needs than are reasonable. Changes to state law that became effective in January 2019 allow HCD to account for present unmet housing needs in addition to future housing needs. Potentially as a result of these statutory changes, some regions received housing needs allocations that are more than double the amount of their previous allocations.

We are aware of two lawsuits that challenge HCD’s process, including one that alleges that HCD did not consider all factors as required by state law. In one lawsuit, the Orange County Council of Governments, which is independent from the larger Southern California Association of Governments, sued HCD, alleging that HCD failed to use the appropriate population forecast, failed to appropriately evaluate household overcrowding and cost burden rates, and used unreasonable vacancy rates. In the other lawsuit, several interested individuals and two nonprofit corporations filed a lawsuit alleging that HCD failed to consider data regarding the relationship between jobs and housing in its assessment for the Association of Bay Area Governments, which is the San Francisco Bay Area council of governments. Both lawsuits are pending final resolution. To avoid interference, we did not review the needs assessments for either of the councils involved in these lawsuits as part of this audit.

The needs assessments affect the planning for housing availability across the State and are an important but sometimes contentious component in addressing California’s housing crisis. Housing availability and affordability has become a key economic issue, as the Legislative Analyst’s Office (LAO) reported in 2019. The LAO noted that the significant shortage of housing, particularly within coastal communities, contributed to higher housing costs for Californians. The LAO also noted that high housing costs increase the State’s poverty rate and, in particular, put low‑income Californians at risk of instability and homelessness. As discussed above, the State’s role in identifying existing and future housing needs to guide the housing planning process is under public scrutiny. Determining accurate, appropriate, and defensible housing needs is a key step in facilitating state and local efforts to plan for housing development.

---

HCD’s Housing Needs Assessment Process Lacks Sufficient Reviews and Support

Key Points

• HCD made several errors when entering data into calculations for its needs assessments, which reduced the amount of housing needs in the needs assessments for two of the three regions we reviewed. HCD does not have a sufficient management review process to ensure that it identifies such errors before finalizing needs assessments. Without effective review processes, HCD may be making similar errors in needs assessments for other councils of governments.
• HCD could not demonstrate that it followed work group recommendations when it considered the balance between jobs and housing, and did not maintain consistency in its consideration of housing destroyed during a state of emergency, when it produced the needs assessments for the three regions we reviewed. In at least one needs assessment, the omission led HCD to understate housing needs by not accounting for units that had been destroyed in a wildfire.
• HCD did not adequately support its adjustment to the needs assessments to address vacancy rates for the councils of governments we reviewed. Despite the significant effect that HCD’s vacancy rate adjustments have on needs assessments, it has not completed a thorough analysis to determine whether it used the most appropriate value in its calculations.
• HCD’s reviews of comparable regions selected by councils of government have been inconsistent because the department does not have a formal process for such reviews. As a result, it did not identify a problematic proposal from a region and inappropriately reduced its needs assessment.

HCD Has Made Errors When Completing Its Needs Assessments Because It Does Not Sufficiently Review and Verify Data It Uses

HCD does not have an adequate review process to ensure that its staff members accurately enter data that it uses in the needs assessments. As Table 1 shows, state law requires HCD to consider a variety of information for its needs assessments for councils of governments, including population projections, housing vacancy rates, and income data. HCD staff members enter the data the department obtains from various sources into a spreadsheet for each council of governments and uses the information to determine the housing needs. However, HCD does not sufficiently review its staff member’s data entries for accuracy. As Figure 3 shows, we noted data entry errors in two of the three assessments we reviewed. We discuss the other issues presented in Figure 3, including an inadequate consideration of the relationship between jobs and housing, in the following section.

Figure 3

HCD’s Errors and Omissions Understated the Needs Assessments for Multiple Regions

One data entry error resulted in a lower, inaccurate number of needed housing units in the Santa Barbara Association’s needs assessment. HCD’s needs assessment letter explained that its overcrowding adjustment relied on Census estimates from five years of survey data. However, HCD had only used Census data from a one‑year estimate when determining the overcrowding adjustment, which is both less accurate and inconsistent with other steps in the calculation that used the five‑year estimates. HCD explained that staff members entered data from the wrong table on the Census website. Had HCD used the five‑year estimates as it intended for this step in its calculation, Santa Barbara’s needs assessment would have included 1,338 more housing units, or about 5 percent more than the inaccurate assessment HCD provided to the Santa Barbara Association.

HCD made a similar error when using Census estimates to adjust the Sacramento Council’s assessment. It had intended to use the 2013–2017 Census vacancy estimate for all the counties within the Sacramento Council, but it mistakenly entered the 2012–2016 estimate for Sacramento County. This error reduced the Sacramento Council’s needs assessment by 2,484 units. Although this number represents a small portion of the region’s overall needs assessment of more than 153,000 units, it still represents homes for individuals and families for which the Sacramento Council needs to plan to accommodate.

Because HCD did not verify the information the Santa Barbara Association submitted for its needs assessment, it made an additional error. HCD incorporates into the needs assessments some information it receives from the councils of governments, such as data on overcrowding. The Santa Barbara Association submitted data on comparable regions’ overcrowding rates using the 2014–2018 Census data, which HCD then incorporated into its overcrowding calculation. However, HCD had intended for its calculation to incorporate 2015–2019 data. Although this particular error was not large, it was in addition to the other errors in the assessments we reviewed, as discussed above. It concerns us that HCD does not have a formal review process to ensure that these important housing needs assessments are as accurate as possible.

We identified these errors, which would be difficult to detect in documentation supporting HCD’s needs assessments, by comparing the data in the needs assessments to the correct source documents. Therefore, we expected that HCD would have a robust process for dedicated reviewers and management to verify that staff members retrieve and enter the correct data in the spreadsheets. However, HCD told us that its primary process for identifying errors in its needs assessments is to send a draft assessment to each council of governments for review rather than to have HCD supervisors or other HCD staff members review the drafts.

HCD’s reliance on the councils of governments for checking the accuracy of the needs assessments is problematic. As we discuss in the Introduction, the needs assessment process can be contentious and draws attention from numerous stakeholders. Therefore, some councils of governments may be reluctant to propose changes or corrections to their needs assessments that increase their own housing needs. In fact, two of the errors we identified inaccurately lowered the needs assessments, but HCD stated that neither the Santa Barbara Association nor the Sacramento Council notified HCD of the errors, and no record we reviewed indicated whether the two councils of governments noticed the errors at all.

When we brought these concerns to HCD’s attention, its deputy director of housing policy development (housing policy deputy) stated that the department plans to conduct and document supervisor reviews of its needs assessments for its next planned round of assessments in 2023. It is crucial that HCD do so to ensure that councils of governments plan for the appropriate amount of housing and to maintain public confidence in the validity of the State’s assessments of local housing needs.

HCD Did Not Demonstrate That It Adequately Considered Certain Factors That State Law Requires for Housing Needs Assessments

HCD did not demonstrate that it adequately considered two factors listed in state law when preparing the three needs assessments we reviewed, which potentially further reduced the reliability of its needs assessments. The law requires HCD to review data and assumptions that councils of governments submit for the factors considered in housing needs assessments, and it allows HCD to make adjustments to the needs assessments after this consideration. HCD may accept or reject the submitted information, and it must issue a written determination on the data assumptions for each factor and the methodology it will use.

Although HCD generally included most of the factors outlined in state law in the three needs assessments we reviewed, it did not adequately demonstrate how it considered two factors: the balance between jobs and housing in the region (jobs/housing balance) and housing lost in emergencies, such as wildfires. The housing policy deputy stated that HCD addresses these factors through its projected household data and other adjustment factors, and currently documents that consideration with an assertion in its final needs assessment that it considered all factors specified in state law.

When we asked HCD about its specific consideration of the jobs/housing factor, HCD indicated that it relied on a work group’s draft analysis of jobs/housing relationships. However, this analysis is outdated and provided limited direction for how the jobs/housing balance would affect needs assessments. The housing policy deputy stated that HCD had studied the jobs/housing balance factor in 2010, 12 years ago. The analysis noted that the inconsistent data available between regions makes regional comparisons of jobs and housing difficult and that statewide standardized employment data are not available for comparison purposes. Although it did not recommend specific adjustments for the jobs/housing balance factor, the 2010 work group indicated that HCD should solicit specific information from councils of governments to address this factor. However, HCD did not specifically request such information from the Sacramento Council, the Santa Barbara Association, or Amador County—the three needs assessments we reviewed—in order to determine those needs assessments.

HCD believes that its other adjustments for different factors also addressed the jobs/housing balance factor. Specifically, HCD asserted that its adjustments to address low vacancy rates, high overcrowding, and high cost burdens address jobs/housing balance issues. However, HCD did not provide an analysis that demonstrated how, or to what extent, these adjustments address the jobs/housing balance. The housing policy deputy also noted the potential for inequitable adjustments for jobs/housing balance between regions because regions receive needs assessments at different times but agreed to review data sources and seek academic perspectives on approaches to account for the jobs/housing balance in the next round of needs assessments. HCD also agreed that as part of its review of the jobs/housing balance factor, it would consider either adding a specific adjustment or modifying its other adjustments, such as increasing the cost burden adjustment, to better account for the factor in the future.

The second factor HCD inadequately considered was housing lost during emergencies. HCD did not consider housing lost during emergencies in a consistent manner across different regions, which led it to understate housing needs in the Santa Barbara Association’s needs assessment. State law requires HCD to consider data and assumptions submitted by a council of governments on housing lost during a state of emergency declared by the Governor if that lost housing has not been rebuilt or replaced at the time of the collection of data for the needs assessment. In 2017 the Governor declared a state of emergency in Santa Barbara and Ventura counties due to the Thomas Fire, which destroyed more than 1,000 housing units and other structures. HCD did not consider the loss of units caused by this wildfire, as required by state law, and did not make an adjustment for this factor in the 2021 Santa Barbara Association needs assessment, as it did in another region, which we discuss below. We believe HCD should have worked with state and county officials to consider this factor in the assessment so that the Santa Barbara Association can plan to address actual housing needs.

HCD’s housing policy deputy explained that HCD believes another factor addresses housing lost to fire emergencies. As we discuss in the Introduction, HCD determines the replacement rate at which each council of governments’ region loses housing units and applies an adjustment in the needs assessment to replace housing. The replacement adjustment reflects the average annual rate of housing loss over the past 10 years that a council of governments needs to replace for units that have been destroyed or demolished, or are no longer inhabitable. The housing policy deputy stated that Finance provides it with information on the rate of housing replacement, such as when there is a fire that requires a building to be replaced. Although HCD considered replacement units in the Santa Barbara Association needs assessment, it did not include a separate consideration for units destroyed in emergencies. HCD’s replacement adjustment identified the average rate that housing is replaced in Santa Barbara County based on 10 years of data from Finance. However, this approach minimized the effect of a wildfire by combining it with normal years of housing losses, resulting in less overall housing than actually needed.

Furthermore, HCD’s approach to the Santa Barbara Association’s declared state of emergency was not consistent with the approach it took in another assessment. Specifically, for the Butte County Association of Governments, HCD worked with county and state officials, including Finance, when it considered and then included an adjustment specifically for housing destroyed in the 2018 Camp Fire, for which the Governor also declared a state of emergency. HCD noted that it included the adjustment for the Butte County Association of Governments because this fire and associated housing loss was particularly large. We expected HCD to consider housing lost in declared emergencies consistently.

HCD needs to thoroughly document its required consideration of each factor because the needs assessment process is complex and can be contentious, drawing significant attention from local governments as well as interest groups. Therefore, it is critical that HCD’s actions increase confidence in the needs assessment process. Although state law permits HCD to determine what adjustments, if any, to make in response to a particular factor, documenting the specific methodology and determination will enhance transparency and public trust. It will also allow HCD to more effectively justify its conclusions to stakeholders and potentially avoid litigation. It is also important that HCD conduct its needs assessments consistently across different regions and in compliance with state requirements, especially when adjusting for sensitive issues such as wildfire disasters.

The Healthy Vacancy Rate HCD Used in Assessments We Reviewed Was Poorly Supported

HCD did not provide adequate support for a critical determination it made about the healthy housing vacancy rate that it used in the three needs assessments we reviewed, raising questions about whether HCD can support the rate in its other assessments. State law requires HCD to consider how councils of governments’ vacancy rates compare with healthy vacancy rates when determining housing needs assessments. As we discuss in the Introduction, state law specifies that a healthy vacancy rate for rental housing should not be less than 5 percent, but it does not specify a healthy vacancy rate for owned housing, allowing HCD to make that determination.

HCD used a 5 percent healthy vacancy rate for the combined rental and ownership markets for two of the councils of governments’ assessments we reviewed. ² HCD used a 4 percent healthy vacancy rate to perform the adjustment for Amador County—a county without a council of governments. HCD explained that it used a lower rate for rural areas because they have a higher proportion of owned housing compared to rental housing and the ownership market typically has less turnover, and thus fewer homes on average will be empty at any given time in rural areas than in the State as a whole. HCD calculated the vacancy rate adjustment by subtracting the region’s overall vacancy rate from the 5 percent healthy vacancy rate. Based on that rate, the vacancy rate adjustment for the Santa Barbara Association resulted in an increase of more than 4,000 housing units to the overall housing needs. Even a 1 percent difference—higher or lower—can make a significant difference in the needs assessment. For example, if HCD had used a 1 percent higher healthy vacancy rate target, the adjustment would have increased by 40 percent, to 5,600 housing units. Therefore, it is important that the rate that HCD uses is adequately supported.

HCD concluded that its choice of a single healthy vacancy rate for the overall market instead of separate rates for owned and rental housing was appropriate. HCD stated that in 2018, for the current round of needs assessments, it began evaluating vacancy rates across the total number of homes available, a change from its previous approach of separating the rental and ownership markets before evaluating vacancy rates in each of them. HCD stated that it changed its approach to reflect the fact that some owned housing becomes rental housing over time. Conversely, a development may be rented for an initial period and then sold to owners after a condominium conversion. However, as shown in Figure 4, the vacancy rates of the two categories are significantly different—ownership vacancy was much lower than rental vacancy over the past 15 years. We are concerned that HCD has not completed a formal analysis to support its claim that a single healthy vacancy rate was appropriate.

Figure 4

HCD Targeted a Vacancy Rate That Is Between Historical Rates for Rented and Owned Housing

When we asked HCD for its support for using the 5 percent healthy vacancy rate in the assessments, it provided only limited information that did not adequately support its assumptions. HCD explained that although it understands that the ownership vacancy rate is somewhat lower than 5 percent, the literature it reviewed indicated that a healthy rental vacancy rate is likely somewhat higher than 5 percent, and it believes the 5 percent is defensible for the combined market. However, HCD did not thoroughly analyze vacancy rates when it began to use this healthy vacancy rate assumption in 2018. HCD provided a summary document from a work group it convened in 2010 that reviewed historical vacancy rates in different regions, but the work group’s summary did not reach a conclusion on a healthy vacancy rate. Instead, the summary referenced information the work group had reviewed, including government reports, and noted a range of vacancy rates among other states that included separate rates for owned and rented housing. Additionally, some of the information was outdated because several of the government reports the summary cited were published in the 1980s. The summary also stated that HCD had used the same healthy vacancy rates—using separate rates for owned and rental housing—since 2006 and may adjust them for current economic conditions.

Despite the large impact of the vacancy rate adjustment on a region’s total needs assessment, HCD has relied on the 5 percent healthy vacancy rate without providing adequate support for its approach. For example, HCD made a vacancy rate adjustment to increase Sacramento’s needs assessment by more than 22,700 units, or nearly 15 percent of the total housing needs. Therefore, we expected HCD to provide sufficient analysis and support for its assumptions underlying the healthy vacancy rate it used in the assessments we reviewed. When HCD does not develop a strong analysis with clear justification for its assumptions, especially those that have significant impact on the size of its final assessments, it risks making adjustments that are not reflective of a region’s true housing needs.

HCD Did Not Identify a Problematic Proposal From a Region and Inappropriately Reduced Its Needs Assessment

HCD did not sufficiently review the regions that councils of governments compared themselves to as part of the needs assessment process. For two factors in its needs assessments, state law requires HCD to consider how a council of governments’ regional data compares to that of other similar regions in the nation. For these factors—overcrowding and cost burden—the law requires councils of governments to provide data from regions they propose as “comparable.” For the cost burden adjustment, state law requires councils to provide data from “healthy” housing markets. State law allows HCD to adjust a council of governments’ needs assessment based on these factors, thus allowing communities to plan for more housing to better address the housing crisis. Under state law, HCD must consider the information a council of governments submits, though it does not have to use that information in its final needs assessment. State law does not provide criteria for the councils of governments to select comparable regions to propose. However, in correspondence to the council of governments we reviewed, HCD recommended that several non‑housing factors—such as population, median income, and jobs per capita—be included for comparison to help guide councils of governments in their selections of comparable, healthy regions.

HCD’s reviews of comparable regions selected by councils of government have been inconsistent because the department does not have a formal process for such reviews. The housing policy deputy explained that HCD reviews the appropriateness of the regions that councils of governments propose as comparable and has rejected a proposal in the past. However, HCD does not have a documented process to guide its evaluation of councils of governments’ proposals to ensure that its reviews are consistent. HCD explained that even though it does provide guidance on what criteria councils of governments could use for their proposals of comparable regions, it has avoided instituting a specific, formal review process because state law specifically allows councils of governments to determine what regions are comparable. However, state law also gives HCD the ability to reject those same proposals. Therefore, we believe it is important for HCD to have a formal process to review the comparable regions that councils of governments propose so it can ensure that it is using this authority consistently for different needs assessments.

The Santa Barbara Association provided HCD with a comparable region proposal that we found problematic. In January 2021, after working with HCD to adjust its comparable region proposal, the Santa Barbara Association provided a memo to HCD explaining that it based its selection of comparable regions on certain categories, such as population, household size, rent‑to‑income ratio, age distribution, and poverty. These criteria resulted in the Santa Barbara Association choosing regions that were likely experiencing housing problems similar to its own region because they also had higher, unhealthy, rates of overcrowding and cost‑burdened households compared to national averages. The use of household sizes and rent‑to‑income ratios to select comparable regions was problematic. For example, the overcrowding rate—reflecting the number of housing units that have more than one person per room in a region—is likely higher in a region with a higher average household size. Similarly, a region with a higher rent‑to‑income ratio is likely to have more households with heavy cost burdens. Higher overcrowding and heavier cost burdens than the national average indicate that those housing markets are not healthy.

HCD accepted the comparable regions the Santa Barbara Association proposed, which likely lowered the needs assessment from what it would have been had HCD used healthy housing markets for one of the adjustments. HCD explained that it views its role as providing guidance to councils of government in their process of selecting comparable regions, rather than being prescriptive. However, our concern is that the Santa Barbara Association specifically used certain criteria that resulted in it selecting unhealthy housing markets, which HCD acknowledges is an approach that has led it to reject other councils’ comparisons. Had HCD compared the Santa Barbara Association to regions with cost burden rates closer to the national average, we estimate that its needs assessment would have increased by 470 housing units to about 25,300, or an increase of 1.9 percent. Without a consistent process to review the criteria that councils of governments propose to identify comparable regions, HCD may be allowing some regions to plan for less housing than they otherwise should.

---

Finance Provides Reasonable Population Projections, but It Has Not Provided Sufficient Support for Its Household Formation Projections

Key Points

• Finance’s population projections are the basis of HCD’s needs assessments, and they are generally accurate. Projections for counties with less than 250,000 residents were less accurate than for counties with more than 1 million residents, but the accuracy of projections has improved over time.
• Finance also creates projections of the number of future households in the State by county. Although HCD uses the household projections in its needs assessments, Finance has not conducted a rigorous analysis to support the household formation rates it uses for the projections.

Finance’s Population Projections Have Generally Been Accurate

The basis of housing needs assessments are population forecasts that Finance produces. State law requires Finance to produce short‑ and long‑range projections of the population, and it does so for the entire State and its counties. To develop its population projections, Finance projects future births, deaths, and migration, or movement into and out of the State, to determine the State’s future population by county. HCD then uses the projections for five to 10 years into the future in its needs assessments, depending on the period the assessment covers. ³ HCD’s needs assessments we reviewed are for eight to 10 years in the future, ranging from 2029 to 2031. To review the accuracy of Finance’s previous population projections and their potential impact on HCD’s needs assessment process, we compared the statewide population projections for 2020 that Finance published in 2011 to Census data for 2020. We found that its projections were overestimated by just 2.7 percent. The variables that affect population estimates, such as the number of deaths, births, and migration, are not constant values and are difficult to predict precisely; therefore, we considered Finance’s statewide projections reasonable.

We also reviewed the process and data that Finance uses to make its projections and found that it is appropriate. Finance has programmed the software that it uses to make projections to identify and remove illogical results and fix errors in the results. Finance staff members also perform reviews of these projections. Staff members compare the projections to previous projections to ensure that there are no unexpected or dramatic changes. Finance also stated that managers review the results before the department provides the data to HCD.

When we reviewed Finance’s county‑level projections over several years, we noted that their accuracy varied. The projections Finance made in 2011 for the 2020 population were less accurate in counties with less than 250,000 residents than in counties with more than 1 million residents. For example, Finance projected that Colusa County’s 2020 population would be nearly 25,000, but the actual population according to the 2020 Census was only about 22,000, a difference of 12 percent. In contrast, Finance projected that Orange County’s 2020 population would be 3.2 million, and the actual 2020 population was 3.19 million, a difference of 0.4 percent. However, we reviewed subsequent projections that Finance published in 2013, 2016, and 2019 of 2020 county populations and found, as would be expected, that its 2019 projections were more accurate.

Finance plans to account for 2020 Census results when making its next population projections in 2023. When we asked Finance about the differences that we identified in its projections compared to Census data, it had already begun reviewing those differences in preparation for its next population projections. In fact, it had identified a series of events and changes that may have affected the accuracy of its projections in specific counties. For example, Finance noted that its projection for Mono County was inaccurate due to population reductions resulting from staffing changes at a military facility in that county. Further, it explained that it overestimated international migration into Imperial County, leading to differences between the Census data and its projection. As a result, Finance told us that it plans to make adjustments in its approach for projections as it incorporates 2020 Census data into its next population projections, which it expects to release in early 2023.

Finance Has Not Adequately Supported Rates It Uses to Develop Household Formation Projections

Finance did not have a rigorous process to support its projections of the number of households in each region, despite the importance of this data in determining a region’s housing needs. One of the factors that HCD’s needs assessments include are the projections of the number of households that Finance expects in future years in communities across the State. Finance estimates the number of expected households by identifying a household formation rate for different age groups in each county. The household formation rate represents the likelihood that individuals in particular age groups will have their own households. HCD applies the rate by age group to the population projections to estimate the number of households that will exist in the future in a region. Because local governments will need to plan housing to accommodate these new households, HCD includes this expected new demand in its needs assessment process.

We expected Finance to use household information in the 2010 Census as its basis for projecting household formation rates, as 2010 data forms the basis of its current set of population projections. ⁴ Finance expects to receive detailed 2020 Census information by county in August or September 2022. It plans to release new population projections, which will include information that accounts for the effects of the COVID‑19 pandemic, in January or February 2023. However, Finance explained that instead it estimated current household formation rates using information from earlier Census data as well as the 2010 Census. Specifically, Finance projects that Californians will be increasingly likely to form their own households in the coming years until household formation rates reach levels seen before 2010. Finance explained that before 2010, more people were willing to live independently than do currently. However, Finance noted the 2010 Census identified a relatively low household formation rate, which may have resulted from cultural, demographic, or economic changes, such as the Great Recession that began in 2007. According to Finance, its household formation rate reflects an assumption that household formation patterns in California will increase over time to pre‑2010 levels—those before that recession, when people were more likely to own homes or take on fewer roommates.

However, Finance did not formally study how Californians would form households. In partnership with HCD in 2014, it solicited advice from some experts participating on the 2015–2025 Statewide Housing Plan Technical and Research Advisory Committee (advisory committee) to guide its decisions on household formation rates. Finance noted that its household formation rates were the result of deliberations among members of the advisory committee. This advisory committee is different from the work group mentioned previously that HCD convened in 2010 that discussed vacancy rates. However, our review of available documentation from the advisory committee found that it did not make any conclusions about household formation rates. The advisory committee also did not provide Finance any formal guidance, analysis, or report on household formation rate trends.

In 2015 and 2016, Finance and HCD staff members reached out to several university professors and other experts from the advisory committee to discuss household formation rates. In a series of emails, staff members from Finance and HCD communicated with experts to discuss factors that may affect household formation rates, such as changes in young adult behavior after the Great Recession and slowing immigration and birth rates. This discussion also reflected concerns about relying on 2010 Census data, because the data reflected conditions during a recession. As part of these conversations, HCD and Finance proposed to the experts several different household rate trends, one of which Finance now uses. Although Finance believes its household formation rates are reasonable, these discussions do not constitute a thorough analysis. Given that this rate is an important component of the household projections that Finance used for multiple years, we expected Finance to better support the assertion that it is using the most appropriate rate. For example, Finance could have documented an analysis of historical household formation trends, a review of academic literature, and its consideration of all factors relevant to household formation rates to demonstrate that its household projections are defensible.

Needs assessments can change significantly depending on the accuracy of Finance’s assumptions. Slight changes to household formation rates, which directly increase or decrease the number of projected households, can change HCD’s needs assessments by thousands of units. For example, if HCD’s needs assessment for the Santa Barbara Association used household formation rates 1 percent lower, the region’s needs assessment would decrease by 17.5 percent, or about 4,350 fewer units of housing. ⁵ The Santa Barbara Association’s current needs assessment calculates the number of projected households using a set of eight household formation rates for different age groups, ranging from 11 percent for residents 15 through 24 years old to 72 percent for residents who are 85 and older. Finance explained that older residents have a higher household formation rate because they are likely to be financially independent and thus live in their own households. Similarly, if the needs assessment used 1 percent higher household formation rates, the needs assessment would increase by as many units.

Finance plans to reevaluate its household formation rates soon. Finance believes the household formation rates it uses are still reasonable because available Census data generally indicated that it was still a reasonable expectation for household formation rates to increase in the future and that it would make sense to wait to formally reevaluate its assumption after detailed 2020 Census data is available. Finance also explained that its assumption that household formation rates will grow over time helps it to avoid projecting that recession‑era economic issues and housing affordability problems will persist and affect household growth indefinitely in the State. However, without a formal comprehensive review of more recent demographic and economic information, Finance cannot adequately assure the public, stakeholders, and HCD that it is providing the most appropriate household formation rates that HCD includes in the critical needs assessment process.

We conducted this performance audit in accordance with generally accepted government auditing standards and under the authority vested in the California State Auditor by Government Code section 8543 et seq. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on the audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives.

Respectfully submitted,

MICHAEL S. TILDEN, CPA
Acting California State Auditor

March 17, 2022

---

Appendix A

HCD Housing Needs Assessments We Reviewed

The chair of the Joint Legislative Audit Committee (Audit Committee) directed the California State Auditor (State Auditor) to conduct an emergency audit to examine HCD’s regional housing needs determination process. We reviewed three of HCD’s regional housing needs assessments: the Sacramento Council, the Santa Barbara Association, and Amador County. We provide those assessments in tables A.1 through A.3 to give context to the findings in our report. As noted in the Introduction, for counties without a council of governments, HCD also provides allocations of housing needs to the county and cities within it. Table A.4 provides the allocation HCD provided to Amador County and the cities within that county. In contrast, the councils of governments provide allocations of housing needs by income category to their member counties and cities.

HCD did not provide consistent details in the three assessments reviewed, and as a result, there are some differences among the assessments we display below. The time covered by the assessments, and the total housing needs that communities must accommodate, vary. HCD does not complete all assessments at the same time and does not always cover the same period because it aligns the needs assessment process with other planning processes, such as regional transportation planning. The total regional housing needs assessment corresponds to the time period displayed either in the assessment header as in the case of the Sacramento Council, or in the population projection.

Table A.1

HCD Regional Housing Needs Assessment for the Sacramento Council

SACRAMENTO COUNCIL: JUNE 30, 2021–AUGUST 31, 2029 (8.2 YEARS)
STEPS TAKEN TO CALCULATE REGIONAL HOUSING NEEDS	AMOUNT
Population: August 31, 2029 (Finance June 30, 2029,    projection adjusted +2 months to August 31, 2029)	2,844,860
– Group Quarters Population	– 57,315
Adjusted Household Population	2,787,545
Projected Households Minus South Lake Tahoe	1,021,005
+ Vacancy Rate Adjustment (2.23%)	22,730
+ Overcrowding Adjustment (0.60%)	6,111
+ Replacement Needs Adjustment (0.50%)	5,105
– Occupied Units Estimated (June 30, 2021)	– 908,396
+ Cost Burden Adjustment	6,957
Sixth Cycle Regional Housing Needs Assessment Total	153,512 Housing Units

Table A.2

HCD Regional Housing Needs Assessment for the Santa Barbara Association

SANTA BARBARA ASSOCIATION: PROJECTION PERIOD (8.6 YEARS)
STEPS TAKEN TO CALCULATE REGIONAL HOUSING NEEDS	AMOUNT
Population: February 15, 2031 (Finance June 30, 2031,    projection adjusted ‑4.5 months to February 15, 2031)	488,190
– Group Quarters Population	– 27,525
Adjusted Household Population	460,665
Projected Households	160,850
+ Vacancy Rate Adjustment (2.51%)	4,030
+ Overcrowding Adjustment (6.44%)	10,359
+ Replacement Needs Adjustment (0.50%)	804
– Occupied Units	– 152,576
+ Cost Burden Adjustment	1,389
Sixth Cycle Regional Housing Needs Assessment Total	24,856 Housing Units

Table A.3

HCD Regional Housing Needs Assessment for Amador County

AMADOR COUNTY: PROJECTION PERIOD (10.9 YEARS)
STEPS TAKEN TO CALCULATE REGIONAL HOUSING NEEDS	AMOUNT
Population: September 15, 2029 (Finance June 30, 2029,    projection adjusted to September 15, 2029)	40,090
– Group Quarters Population	– 4,405
Adjusted Household Population	35,685
Projected Households	15,330
+ Vacancy Rate Adjustment (0.04%)	6
+ Overcrowding Adjustment (0%)	0
+ Replacement Needs Adjustment (0.50%)	68
– Occupied Units	– 14,697
+ Cost Burden Adjustment	34
Sixth Cycle Regional Housing Needs Assessment Total	741 Housing Units

Table A.4

HCD Regional Housing Needs Assessment for Amador County

	REGIONAL HOUSING NEEDS ALLOCATION BY INCOME CATEGORY
JURISDICTION	VERY LOW	LOW	MODERATE	ABOVE MODERATE	TOTAL
Amador County Total	189	123	140	289	741
Amador	1	1	1	2	5
Ione	30	20	25	42	117
Jackson0	27	23	24	64	138
Plymouth	7	5	5	13	30
Sutter Creek	15	12	13	34	74
Unincorporated Amador County	109	62	72	134	377

---

Appendix B

Scope and Methodology

The Audit Committee directed the State Auditor in October 2021 to conduct an emergency audit to examine the regional housing needs determination process. The audit was approved under Joint Legislative Audit Committee Rule 17. Recognizing that Rule 17’s cost limitations prevented us from satisfying all objectives of the emergency audit, we focused our work on the first three objectives contained in the emergency audit request. The table below lists those objectives and the methods we used to address them.

Audit Objectives and the Methods Used to Address Them

AUDIT OBJECTIVE		METHOD
1	Review and evaluate the laws, rules, and regulations significant to the audit objectives.	Reviewed relevant laws, rules, regulations, policies, and procedures related to the housing needs assessment process.
2	Assess Finance’s process for developing population projections used by HCD. Determine what changes Finance made to its projections in response to economic and demographic changes caused by the pandemic as well as new Census information. Evaluate historical accuracy of Finance’s population projections	Reviewed Finance’s calculation process for its most recent set of projections and assessed the reasonableness of its process and the information Finance uses to generate its projections. Assessed Finance’s planned modifications to future projections based on COVID‑19 impacts and found them to be reasonable. Finance intends to update its projections in January or February 2023 to take into account recent Census data that reflects reduced births and increased deaths due to the pandemic in 2020 and early 2021. Compared Finance’s past population projections to 2020 Census data to assess their accuracy.
3	Evaluate HCD’s process for developing regional housing needs determinations to ascertain whether it complies with state law and results in appropriate calculations. Assess whether HCD properly used vacancy rates for rental markets and for the entire housing market.	Reviewed the process HCD used to create three needs assessments for the Sacramento Council, the Santa Barbara Association, and Amador County, and determined which factors listed in state law it considered, and whether its consideration was appropriate. For the same three assessments, which HCD completed after changes to state law in 2018, reviewed each adjustment HCD made in the assessments and determined the relative impact of the adjustments on the overall assessment. For the three assessments we reviewed, assessed HCD’s support for the 5 percent healthy vacancy rate it uses for the overall housing market, including reviewing available historical information and economic research.

Source: Audit workpapers.