The Impact of the Subsidized Rents on the City Tenure of Low-Income Renters

According to the U.S. Census, 58 percent of the 292,000 occupied housing units in the District of Columbia in 2019 were rental units. Thus, rental housing is the preferred housing type for most in the city and plays an important role in the city’s overall housing market. However, rapidly rising costs of even this type of housing (relative to home ownership) is a growing concern for many. In this context, one may suppose the high income of many renters would be positively correlated with their city tenure (i.e. presumably a greater ability to pay). But, this appears not be the case. By way of a closer examination of the role of different rent levels on the city tenure of renters, we find the subsidized monthly rents for respective residents are a significant factor contributing to their relative longer average city tenure, and the very high monthly rents of Class A residents are a significant factor contributing to their shorter average city tenure.

To better understand the role of housing costs in the city’s growth and development, the Office of Revenue Analysis (ORA) conducted a study (see here) to see if different rent levels had any effect on the length of city tenure (i.e. the years in the city) of apartment tenants in the city. This study examined this issue from two perspectives. First, the study assessed the effect of subsidized apartment rents (via Housing Production Trust Fund housing) on respective tenants’ tenure in the city relative to apartment tenants with market rents in comparable Class B and C buildings. And second, the study assessed the effect of relatively high apartment rents (i.e. in large Class A multifamily buildings) on the respective tenant’s tenure in the city relative to apartment tenants with market rents in comparable Class B buildings. Class A buildings are the newest and highest quality buildings built within the last 15 years with top amenities. Class B buildings are generally older buildings and more likely to have lower income tenants. And, Class C buildings are the oldest buildings in the market and tend to be in need of major renovation but have the lowest rental rates.

The Housing Production Trust Fund

The Housing Production Trust Fund (HPTF) is a revenue fund administered by the D.C. Department of Housing and Community Development (DHCD). The HPTF provides gap financing for new and renovated residential projects meant to be affordable for low- to moderate-income households. The goal of the HPTF program is to make sure that “every resident in the District can afford a place to call home.” Since 2001, more than 9,000 affordable housing units have been produced using the Fund’s resources. At the end of fiscal year 2019 the HPTF had an available balance of $142.9 million. [Annually, 15% of deed recordation and transfer revenue from property transactions are dedicated to the Fund. In 2019, that amount was $77.2 million and in 2020, $68.1 million.]

Segmenting the District’s Apartment Rental Market

We compared city tenure of income tax filers who lived in HPTF multifamily buildings for at least one year between 2008 and 2012 to income tax filers who resided solely in market-rate (i.e. non-subsidized rent) Class B or Class C buildings during the same years and in the same neighborhoods. And, in efforts to be more comprehensive in understanding the city’s rental market, we compared city tenures of income tax filers that were residents of large Class A multifamily buildings which have higher rents for at least one year between 2008 and 2012 to income tax filers that were solely residents of market-rate Class B buildings for the same time period also in the same neighborhoods.

The Role of Apartment Rent Levels

Figure 1 shows that income tax filers who only lived in market rate Class B large multifamily buildings, and controlling for other factors, remained city residents an average of six years. However, income tax filers who lived in an HPTF building tended to remain city residents for almost 2 years longer, and filers who lived in large Class A multifamily buildings tended to remain city residents a little less than 5 years.

Figure 1

The study found that HPTF residents (relative to market rate Class B rental unit residents) followed by head of household tax filers appear to be the strongest factors identified contributing to longer city tenures as shown in Figure 2. In contrast, Figure 3 shows that for income tax filers who resided in Class A rental units in large multifamily buildings, filing as head of household (relative to filing as married or single) was the strongest factor identified contributing to longer city tenures while residing in a Class A rental unit was also strongly correlated with shorter city tenures. In the models, “being a HPTF resident” is our proxy for having low housing costs (i.e. subsidized monthly rents), “being a Class A resident” is a proxy for enduring some of the city’s highest monthly rents, and “being a Class B resident” is our reference group.

Figure 2

Figure 3

The Role of Income

The study also found that the average income of residents of HPTF rental housing in 2016 was $22,978, while the average income of residents of Class A rental housing was more than 3 times higher at $75,326. However, the effect of average income growth on increasing city tenure was 3.5 times stronger for HPTF residents than Class A residents. This might suggest that, as HPTF residents experience income growth, their subsidized rents may become even more valuable to them when considering the other residential alternatives, including migrating out of the city. Hence, income growth for low income earners appears to be positively correlated with their longer city tenures but has little correlation with longer city tenures by Class A tenants.

Tenants with the Lowest (Highest) Incomes have the Longest (Shortest) Tenures

Because of the city’s high housing costs, one may suppose that high income would be positively correlated with city tenure (i.e. presumably a greater ability to pay). But, the above findings indicate the opposite. We find that the low monthly rents for HPTF residents is a significant factor contributing to their longer average city tenure, and the very high monthly rents of Class A residents are a significant factor contributing to their shorter average city tenure. (While it is probable that at least some outmigrant Class A residents move into homeownership outside the city, such data were not analyzed for this study.) And while there are many factors that influence city tenure, this study focused primarily on the role of rent levels on the city tenure of city residents.

HPTF: Helping to Retain an Economically Diverse City

Since 2006, the District of Columbia has reversed the trend of decline and has experienced significant population growth. However, there continues to be a simultaneous and persistent pattern of out-migration of residents particularly at the lower levels of income. This study finds the level of apartment rents faced by the city renters appears to influence their length of tenure in the city. It also appears that HPTF supported housing is providing an increasing number of affordable housing units to low-to moderate- income residents, which, in turn, is helping to counter the more general pattern of out-migration of low-income residents caused, in part, by the lack of affordable housing and gentrification.

What is this data?

This study utilized administrative tax data that includes District of Columbia individual income tax data for 2001 to 2016, property tax data for 2001 to 2016, and federal individual income tax data for city residents for years 2006 to 2016. The tenure of tenants is based on the cumulative number of years tenants filed their income taxes with the city from their particular residences. Also, HPTF housing data was collected from the DHCD Open Data database and HPTF Annual Reports.

The key model focused on income tax filers that were residents of HPTF multifamily buildings for at least one year between 2008 and 2012 (the treatment group). Accordingly, income tax filers who resided solely in market-rate (i.e. non-subsidized rent) Class B or Class C buildings between 2008 and 2012 are the control group. The use of tenants’ income tax data for years 2001 to 2007 and 2013 to 2016 helped to more accurately identify the start and ending year of city residency and tenure of all tax filers in the study. All tax filers were between the ages of 22 and 70 when they first moved into any of the buildings in the study.

An Evaluation of the District’s Tax Increment Financing: Is it a net fiscal gain to the District?

Tax Increment Financing in the District

Tax Increment Financing (TIF) is an economic development policy tool used by state and local governments to stimulate economic development in a specific area. Its popularity stems from the notion that a TIF project is self-financing, that is, that the tax revenue generated by the project covers the bond payments for the publicly subsidized loan needed to implement the project. The District of Columbia implemented its first TIF project in 2002 and that was followed by seven additional large TIF projects. (There are some smaller TIF projects but these eight projects in our analysis are the largest, most important ones.) Our recent study (See here) is a retrospective evaluation of the economic and fiscal performance of these District TIF projects.  The study’s objective was to answer two questions: 1) does each project produce a positive net fiscal gain for the city; and 2) does the entire TIF program produce a positive net fiscal gain for the city?

The study of the eight projects finds that five were indeed self-financing, while three were not (see Table 1). We also found the net tax revenue from the five positive projects (property and sales taxes minus debt service) covers the shortfalls in the three lagging projects so that the District’s TIF program in the aggregate appears to be a net fiscal gain to the city.

A Review of the District’s TIF Projects

By and large, tax increment financing in the District of Columbia is generally used to help produce large but unconventional development projects in specific locations of the city that otherwise, arguably, would not happen. The actual amount of a project’s TIF subsidy is the principal amount plus interest of the TIF debt service. Typically, this subsidy directly finances all or a portion of the total development costs of the project. Theoretically, the TIF subsidy is justified as the amount needed to overcome some stated economic impediment for example, environmental damage, keeping a site location from achieving its highest and best use. But in the District of Columbia, it appears that TIFs are applied for a slightly different reason.

The District of Columbia is a relatively small city with over half of its land area prohibited from being developed by the private markets, largely because of the federal presence. The city also has a vibrant economy that generates steady growth in jobs and population. The result is that development of one sort or the other is happening in almost all areas of the city. As such, TIFs in the District are used to facilitate development projects in specific neighborhoods that, possibly, might not see that exact type of development, usually to achieve important socio-economic goals beyond just job and income growth. For example, the Gallery Place and Mandarin sites may have likely been developed as predominately Class A office space without TIF. But Gallery Place is now the entertainment and retail hub of the city’s central business district, and the Mandarin is a 4-star international hotel that was the first to bring upscale hotel and retail activity to the Southwest waterfront. DC USA may have likely been a predominately residential development without TIF.  But, it was the first new large-scale retail complex in a residential neighborhood anchored by one of the nation’s leading national retail chains. And the Convention Center Hotel site may have also likely been a predominately Class A office space or even a hotel development (with only a fraction of hotel rooms, thus precluding it from being classified and marketed as a convention center hotel). Instead, it is now the city’s largest hotel with 1,175 rooms and an underground concourse connecting it to the Convention Center (See here).  So, we can view TIF subsidies in the District of Columbia as the amount needed to overcome the additional project costs above and beyond the costs of a project that would be the highest and best use of land in that location. The subsidy is justified as a means of achieving important socio-economic goals for the city (e.g., grocery stores to eliminate food deserts in certain neighborhoods, broadening the market for business travel to the District, and increasing affordable housing).

The District of Columbia implemented its inaugural TIF project in 2002, and that was followed by seven additional large TIF projects up until 2010. District TIF projects have been used to facilitate retail, residential, hotel, and other mixed-use developments in the city (Table 1).

Measuring the benefits: TIF vs. A privately financed alternative project

One of the key assumptions in the analysis conducted by many state and local governments to decide whether to greenlight a potential TIF project is that the location site for the TIF project has prohibitively high economic costs that precludes a purely private sector financed development of the project (the so-called “but for” test).

The model used for this analysis does not make such an assumption. Our model assumes each actual TIF location would eventually be developed by the private sector absent a TIF but as a project much more in line with the conventional economic and social characteristics of the existing neighborhood (i.e., less investment risk).

We assume the privately financed alternative for each project starts sometime between the TIF start date and 2019 and that the 2019 real property value of the alternative project is 75 percent of the TIF project’s real property value in 2019.[1]

For example, the property value of the Gallery Place TIF project grew from $6.9 million in 2002 to $596.6 million in 2019.  We assume the property value of the counterfactual at the Gallery place location grows such that it reaches 75 percent of the Gallery Place project’s real property value (that is, $447.2 million) in 2019 but for the public financial support.

Earlier TIF projects were self-financing, but later ones were not

We use the model to evaluate each TIF project and present the results in Table 2. The table shows that the Gallery Place project achieved a positive cash flow in year 4 of the 25-year debt service schedule and reached its breakeven point in year 8. The table also shows that when we divide total cumulative net tax revenue for the 25 years period by the TIF bond amount (in 2019 dollars), the District is estimated to achieve a 67 percent return on investment (ROI) in year 25 of the debt payment schedule (adjusted for inflation). Also, five years after the project was delivered to the market, the TIF bond amount was 21 percent of the project site’s total property value.

When the estimated ROI in year 25 of each project’s debt service schedule is greater than zero for any project, we conclude the project will ultimately be a net positive fiscal gain for the city. Consequently, the District’s first five TIF projects produced a net positive cash flow, a financial breakeven point (all within eight years), and ultimately a positive fiscal gain (ROI) for the District. However, the Capper Carrollsburg, Convention Center Hotel and Rhode Island Row projects are not expected to produce a positive ROI over the life of the TIF loan as indicated by the negative ROI in column 5.

Based on a closer examination of the model results and the actual specifics of each project, it appears that the first five TIF projects achieved sound financial standing and an early breakeven point primarily because the total debt for each project was relatively low and the net tax revenue generated at each site was relatively high. In contrast, despite their important socio-economic roles, the last three projects (Capper Carrollsburg, which subsidizes senior housing; Rhode Island Row, which subsidizes affordable housing; and the Convention Center Hotel, which was developed to land more and larger annual conventions to fully utilize the Washington Convention Center), are not expected to generate enough net tax revenue over 25 years to service the TIF debt. (See study here)

What is the net fiscal impact of all the District’s TIF program so far?

When we aggregate all eight TIF projects between the date of the first TIF bond issuance (2002) and when the model estimates that the last debt service payment is due (2034), we can calculate the annual total debt service (including interest) and the total annual net tax generated at each TIF site. We find that the entire District’s TIF program achieves a positive cash flow starting in 2005. The model results indicate that the Gallery Place and Mandarin Hotel projects are the primary sources of excess net tax revenue for the program and, hence, are the source of cross subsidizing the seemingly uneconomic projects of Capper Carrollsburg, Convention Center Hotel, and Rhode Island Row.

The District’s TIF fiscal balancing act

According to the analysis, five TIF projects generated a net fiscal gain, and 3 did not. We conclude that while the financials of each new potential TIF in the future must continue to be highly scrutinized, policy makers should also remain cognizant of the financial health of the District’s entire TIF portfolio. The study finds that, to date, several large projects with high ROIs are cross subsidizing those seemingly uneconomic projects. The cross subsidy enables the “uneconomic” projects to achieve very important socio-economic goals for the District (e.g. affordable housing and broadening the market for business travel to the District) without jeopardizing the District’s overall fiscal health.

What is this data?

The model used in this study is designed to determine if each TIF project produced or is likely to produce enough incremental tax revenue to cover its debt service. We define incremental tax revenue as the tax revenue from the TIF project in excess of tax revenue from a comparable (a counterfactual) but totally privately financed project in that location. The model assumes a 25-year loan (i.e. TIF) at a 6 percent interest rate applied to each TIF project but for the actual TIF loan amount for each project. The interest rate for fixed rate debts range from 4 to 7.5 percent, and the actual median interest rate for the seven projects was 6.1 percent.

This analysis is based on annual real property assessment values and real property tax collections for years 2002 through 2019 for each TIF project. The analysis for three projects (Gallery Place, DC USA (Target) and Rhode Island Row) include retail sales activity (sales taxes), and four projects (the Mandarin, Capitol Hill Towers, Embassy Suites, and Convention Center Hotels) include hotel sales activity (hotel taxes). In the model, that actual property, retail and hotel taxes generated at each site is also used to finance each respective project’s TIF debt.

[1] 1 Given that most TIF projects are in prime locations and the city has experience significant property development in nearly all areas of the city, we assume the estimated 2019 real property valuation of each counterfactual would range between 50 and 100 percent of the actual 2019 TIF property valuation. Hence, we assume an average 75 percent 2019 real property valuation of each counterfactual in the model.

DC’s Property Sales Market: Trends and Fluctuations – 2000 to 2020

An Overview of the City’s Property Market

Between 2000 and 2019, the District of Columbia property market experienced remarkable growth. According to annual District property assessment data, the total value of all taxable property grew from $46.6 billion in 2000 to $224.4 billion (382.4 percent higher) in 2019 (Figure 1). The value of the residential sector grew at an annual average rate of 9.5 percent over the period, and the commercial sector grew at a slightly slower annual average rate of 8.7 percent. The higher growth rate for the residential sector caused the share of all residential value to increase to 56.5 percent in 2019 and the commercial sector share to decline to 43.5 percent (Figure 2).

Whereas most of the value of all taxable real property is attributed to the residential sector, most of the value of property sales subject to total deed taxation is commercial property. Based on annual deed tax data from the District Recorder of Deeds, we estimate that the total property value subject to total deed taxation grew from $4.1 billion in 2000 to $17.4 billion in 2019 (Figure 3). The value of the residential property sold grew at an annual average rate of 7.3 percent over the period, and the value of the commercial property grew at an annual average rate of 8.7 percent. And while there are many factors (nationally and locally) that contributed to the rapid appreciation of property prices and values, the 10-year treasury rate also fell from 6.05 percent in 2000 to 2.05 percent in 2019. And in contrast to the rapid growth in property values, the national consumer price index only grew at an annual average rate of 2.1 percent between 2000 and 2019.

Additionally, the faster growth of property sale values in the commercial sector caused the share of all commercial value to increase to 51.1 percent in 2019 and the residential sector share to decline to 48.9 percent (Figure 4).

Figure 5 shows that commercial property sold as a share of all commercial property in the city was 13.4 percent in 2001, and residential property sold as a share of all residential property in the city was 13.4 percent also in 2001. The slight but general downward trend in the ratios over the years in the figure reflect the robust annual growth on the total value of both sectors since 2000. In 2005, a record high of 23.2 percent of the total commercial property value in the city was sold. This reflects the brief period of very rapid expansion in commercial office space in the city prior the Great Recession. A record of 50 investment grade office buildings were sold in 2005 (Delta Associates). 

Annual Fluctuations in the Property Sales Market

The above figures may suggest that the property sales market has grown in a relatively smooth upward trending fashion, but that is not the case. Figure 6 shows that more recently the Great Recession caused deed recordation taxes to decline 35.4 percent in 2009. It appears that major broad-based cuts to federal spending (or threatened cuts) in 2012, 2013 and 2015 contributed to deed taxes declining significantly in those years.  The figure also shows that the current COVID-19 pandemic (and resultant national recession) also caused deed tax collections to decline 23.6 percent in 2020. (Coincidentally, the District increased its deed recordation and transfer tax rates from 1.45% to 2.50% for commercial properties valued at $2 million or higher in October 2019. We estimate absent that tax rate increase, deed tax collections would have declined approximately 38 percent in 2020 compared to 2019, making it the largest decline in deed tax collections in over 20 years.)

Figure 7 shows the annual percent change in deed tax collections by class.  In years of major decline, there tends to be a greater reduction in commercial sales activity than in residential sales activity. The greater contraction in commercial sales activity caused the value of all residential sales to account for 50 percent or more of all value sold in those same years (Figure 8). (Interestingly, the commercial office sales activity in the city began to decline in 2006, three years before the Great Recession of 2009. Hence, this sector may at times be a leading indicator of the broader economy and not just a lagging indicator.) But more generally, figures 7 and 8 suggest that annual residential property sales are the bedrock of annual deed tax collections, and the level of sales or notably lack of sales of commercial property causes the greatest declines/swings in annual deed tax collections activity.

Residential Property Sales in 2020

There were 9,270 residential homes sold in 2019 but only 8,469 (8.6 percent less) sold in 2020. The lower number of homes sold in 2020 is largely attributed to the 23.7 percent drop in single-family attached homes (Figure 9). Interestingly, neither of these three subsectors of the residential market experienced a decline in median sale prices in 2020 (Figure 10).

In addition to single family home sales, the residential property sales sector also includes the sale of multifamily (rental) properties. When we examined the sales of the largest multifamily properties in years 2014 to 2020, we find that there were fewer sales in 2020 than in 2018 and 2019. While this subsector tended to produce an average of $594 million in transaction volume in years 2014 to 2019, only 35 percent of that average was generated in 2020. This appears to be a major factor in the 27 percent decline in deed tax collections from the residential sector in 2020 as shown in Figure 7. Consequently, the total value of residential property sales as share of the total value of all residential property value in city dropped from 6.0 percent in 2019 to 4.3 percent in 2020. The properties shown in Figures 11 and 12 are large Class A & B multifamily residential properties with more than 100 rental units that were built or renovated after 2000. The average sale price for the over 2 million square feet of large multifamily building space sold in 2019 was $373 per square foot (CoStar).

Commercial Office Building Sales in 2020

The city’s commercial sector experienced a greater decline in deed tax activity than the residential sector in 2020. In the large commercial office building subsector, there were also fewer sales and lower transaction volume in 2020 (Figure 13). Although the average annual transaction volume in this subsector for years 2014 to 2019 was about $4 billion, 2020 saw only about 42 percent of that average amount (Figure 14). Assuming 2020 would have been relatively similar to the prior years, this suggests that approximately $2.3 billion in sales transactions did not occur in 2020, which is likely a major factor in the 42.7 percent decline in deed tax collections in the commercial sector in 2020 (Figure 7). Consequently, the total value of commercial property sales as share of the total value of all commercial property value in city dropped from 8.9 percent in 2019 to 5.0 percent in 2020. The average sale price for the over 7 million square feet of large commercial office space sold in 2019 was $516 per square foot (Delta Associates).

An Interpretation of Recent Trends and Fluctuations

Over the past 21 years, DC’s property market experienced not only remarkable growth but also major market fluctuations.  The greatest annual fluctuations appear to be correlated with national recessions, major broad-based cuts (or threatened cuts) to federal spending and the COVID-19 pandemic.  From the perspective of annual deed tax collections, it appears that these national economic shocks took a greater toll on the city’s large commercial office building sales sector than on the city’s residential sales sector. Also, the years in which such shocks occurred were promptly followed by strong rebounds in deed tax activity in both sectors. This suggest that these shocks caused a significant slowdown in sales activity (or even a postponement of a considerable number of sale transactions) for that year and maybe the following year with a relatively prompt return to more normal sales levels resuming shortly thereafter, particularly in the large office building subsector. 

The residential and commercial sectors of the city property sales market each account for about half of all taxable property value sold on an annual basis. The residential sector tends to grow healthily and is the relatively less volatile sector of the city’s property market. Hence, it can be considered the bedrock of annual deed tax collections. The commercial sector, on the other hand, has been growing faster on average and appears to be responsible for most of the volatility in annual deed tax collections activity.