BLOG POSTS ARE PRELIMINARY MATERIALS CIRCULATED TO STIMULATE DISCUSSION AND CRITICAL COMMENT. THE VIEWS EXPRESSED ARE THOSE OF THE INDIVIDUAL AUTHORS. WHILE BLOG POSTS BENEFIT FROM ACTIVE UHERO DISCUSSION, THEY HAVE NOT UNDERGONE FORMAL ACADEMIC PEER REVIEW.
By Justin Tyndall
Honolulu faces an acute housing shortage, yet tens of thousands of homes sit empty. Bill 46 would create a new property tax that applies to empty homes. The policy aims to achieve two key objectives: (1) expand the supply of housing for local residents, and (2) generate revenue for the county. Using current property tax records for Honolulu, this blog provides estimates of the possible magnitude of these two impacts. Under any reasonable assumptions, Bill 46 would either significantly expand the available housing supply, significantly raise county revenue, or both.
The current proposal would see a 3% property tax surcharge on properties that are vacant for more than 6 months in a year. The rate would be phased in, with 1% assessed the first year the property is vacant, 2% the following year, and 3% for years thereafter. The bill provides a list of 15 exemptions. Properties where the owner is actively trying to sell or rent are exempt, as well as properties undergoing renovations, or those affected by medical issues, military service, or other specific extenuating circumstances. Homeowners would need to affirmatively declare that their property is not vacant each year to avoid the tax.
The bill emulates a policy enacted in Vancouver, Canada, which also assesses a 3% surcharge for properties left vacant. In Vancouver, the policy coincided with a 54% reduction in the city’s vacancy rate. In 2023, the Vancouver Empty Homes Tax raised $47 million CAD ($35 million USD) from properties that remained vacant. Some of this revenue was directed into local housing programs, while some entered the city’s general revenue.
To predict the impacts of Bill 46 in Honolulu, there are three major sources of uncertainty. First, we don’t know the true vacancy rate in Honolulu. Second, we don’t know how willing empty home owners will be to convert their units. Third, we don’t know how long units are typically kept vacant.
The text of Bill 46 assumes that 9.2% of homes in Honolulu are currently vacant. The number is based on 2020 U.S. Census estimates. However, about a quarter of these vacancies are from homes that are listed for sale or rent, which would be exempt from the tax. This means roughly 6.9% of housing units would actually be subject to the tax. More recent 2023 data from the American Community Survey (ACS) estimated a 10.0% vacancy rate, or 7.5% when units for sale or rent are excluded. The map below shows the approximate locations of the 26,000 vacant homes in Honolulu that would be subject to the tax. The exact locations of vacant homes are unknown. The locations shown on the map are approximations based on census tract-level ACS survey data. Vacant units appear in all residential areas but are more concentrated in areas popular with second-home owners, such as Waikiki.
Approximate Locations of Vacant Homes
⬤ = 1 vacant home
Locations are approximate and based on census tract-level estimates of vacant units.
The US Census estimates are imperfect and based on who responds to Census questionnaires. For example, if a household fails to respond to Census workers, their house may be mistakenly counted as vacant. Or, if a seasonal resident responds to a Census questionnaire, the home could be considered occupied, even if it sits empty most of the year.
The misclassification of vacation rentals as vacant homes could also inflate the Census estimate. A revision to Bill 46 clarified that the policy would not apply to legally operating vacation rentals. Census surveyors are meant to exclude “transient accommodations,” such as vacation rentals, from the enumeration of housing units, meaning these units should not contribute to the vacancy count. However, because vacation rental properties are unlikely to respond to a Census questionnaire, it’s conceivable that some vacation rentals could be misclassified as vacant. In census tracts that contain a county-defined “resort zone,” where vacation rentals are allowed to operate, the ACS vacancy rate among units not for sale or rent is reported as 27%. Many of these vacant units are second homes, though some could be misclassified vacation rentals. However, vacation rentals only comprise about 2% of Honolulu’s total housing stock, so they are unlikely to significantly affect the estimated vacancy rate. Excluding resort zones and units for sale or rent, the Honolulu vacancy rate is roughly 6%.
Taken together, the various census data points suggest the tax would apply to 6-7% of the housing stock. A secondary benefit of implementing an empty homes tax would be to reveal the true vacancy rate in Honolulu by requiring owners to report and substantiate occupancy status.
A second source of uncertainty is the share of empty home owners who would be willing to convert their unit rather than pay the tax. The tax is sufficiently large that owners would need to seriously consider whether they are willing to pay a 3% tax every year, on top of existing property tax. For a $1,000,000 property, the bill would be $30,000 each year. High-wealth owners may still prefer to keep their units empty or for occasional personal use, but doing so would carry a large cost. In Vancouver, 54% of homes switched from vacant to occupied once the tax was implemented. However, it’s not obvious that this drop should be entirely attributed to the tax. Other policies and larger economic factors could have contributed to the change in Vancouver, but the observed change represents a plausible market response.
The choice to phase in the tax based on how long the property is kept vacant helps reduce the burden on owners who leave a home empty for a transitional period. The phase-in period is not part of the Vancouver policy, which assesses the full 3% immediately. The phase-in may reduce the impact of the policy. ACS data indicate that the majority of vacant units are left vacant for only one year or less. Among units that are not for sale or rent, roughly 80% have been vacant for one year or less, 9% were vacant for 1-2 years, and 11% were vacant for more than 2 years. These estimates might undercount the share of long-term vacant units with respect to the tax because the tax would only consider a vacancy period interrupted if the occupancy period was more than 6 months. Using the ACS estimate, the tax rate faced by the average vacant home owner is only 1.3%.
The tool below allows users to input their own assumptions about vacancy rates, homeowner behavior and vacancy duration to visualize potential policy outcomes. Hovering over different combinations of the vacancy rate and willingness of owners to convert their units reveals different impacts. People may have different opinions about the true vacancy rate, the likelihood a vacant homeowner would convert their unit or the typical duration of vacancy. The tool allows you to view possible policy outcomes for whatever you think the true numbers are.
Housing Supply and County Revenue Calculator
The value of the average vacant home is calculated as the average across all non-owner occupied, non-resort zone properties, weighted by the likelihood the property is vacant based on local Census estimates. The method estimates the average vacant home has a taxable value of $925,000, slightly higher than the average across all residential properties ($874,000). The gray dashed lines indicate Honolulu’s estimated vacancy rate based on Census data and the conversion rate that coincided with Vancouver’s Empty Homes Tax.
The higher the true vacancy rate is, the more impactful the policy will be in terms of both housing supply and county revenue. If more owners prefer to simply pay the new tax and leave their property vacant, the policy will raise significant revenue, but contribute few units of new housing supply. Alternatively, if owners fill their units, the revenue effects would be smaller, while the housing supply effects would be sizable.
In Vancouver, the implementation of the 3% Empty Homes Tax was accompanied by a 54% decline in vacant units. If we assume 7% of Honolulu homes are vacant under the tax, and we assume Honolulu empty home owners will respond similarly to those in Vancouver, the policy would free up 14,000 housing units and the county would take in an additional $144 million each year. If the new local housing supply resulted in a decrease in home values, the county revenue windfall would be proportionally reduced. Given the unknowns, the actual outcomes could vary significantly. Under other reasonable assumptions, the policy might free up anywhere from 4,000 to 20,000 units of housing and could raise annual revenue anywhere from $50 million to $400 million.
The potential housing supply and revenue effects are both significant. An increase in 14,000 housing units amounts to a 4% increase in Honolulu County’s total housing stock. In recent years, Honolulu has only expanded its housing supply by roughly 2,500 units per year, so this change represents several years of new housing supply. On the revenue side, $144 million represents about 4% of Honolulu’s entire county operating budget. This estimate assumes the average unit is assessed at a 1.3% rate. If typical vacancy durations are longer than the ACS data suggest, the tax would raise more revenue. If we assume the average was 2%, the policy would raise $222 million. Whether the funds are used to build additional housing units or provide other government services, the amount of new revenue would be substantial. The empty home tax would raise this revenue disproportionately from second-home owners and out-of-state residents, making the tax relatively progressive and reducing the burden on locals.
Because the surcharge is phased in based on how long the property has been vacant, the policy is less costly to owners compared to the Vancouver Empty Homes Tax, which assessed 3% immediately. This difference could lead to a lower conversion rate in Honolulu, which would mean fewer than 14,000 homes are made available. However, a lower conversion rate would result in more county revenue.
In its current form, Bill 46 applies only to empty housing units and does not apply to vacant land in residentially zoned areas. Purchasing residential land and choosing to leave it vacant also reduces housing supply. An empty lot is assessed at a lower value than if it had a home on it, and therefore enjoys a lower property tax. This incentivizes owners of empty lots to leave them empty, which deprives the county of both housing and revenue. Taxing empty land would incentivize the landowners to quickly build housing to avoid the tax. According to property tax records, Honolulu has more than 2,000 residential lots sitting empty, each with the potential to provide multiple units of housing. Extending the tax to cover empty lots in addition to housing units could provide additional housing and revenue.
Bill 46 would discourage investors from holding vacant homes on O’ahu specifically, but this could potentially redirect investment pressures to the Neighbor Islands, as O’ahu would become relatively less attractive for second-home buyers. Honolulu’s experience with implementing an empty homes tax may hold lessons for the Neighbor Islands, and other jurisdictions struggling to expand housing supply for locals.