Evolution of the New Market Tax Credit

1 Introduction

Governments frequently implement place-based policies, such as Opportunity Zones in the United States, the Local Enterprise Growth Initiative in the United Kingdom, or Law 488 in Italy, to incentivize investment and improve economic conditions in impoverished areas. The New Market Tax Credit (NMTC) attempts to increase the flow of financial capital to businesses in census tracts which meet certain criteria, primarily sufficiently low income or high rates of poverty. The NMTC program has provided $25 billion worth of tax credits for financial investment in designated communities since its inception in 2000.

In this paper I characterize how investments from the program and claimants of the NMTC have changed over time. I focus on the targeting of credits to rural versus urban areas, the role of real estate projects, and characteristics of NMTC claimants. The share of real estate projects has declined by over 30 percentage points and the share of projects in metropolitan areas has declined by over 20 percentage points since 2006. I document that NMTC claimants have higher levels of income and assets than the average taxpayer. I also illustrate the increase in the relative importance of the NMTC as it has become a higher share of general business credits for NMTC claimants. By 2018, the NMTC had risen to over 45% of the general business credits for NMTC claimants. Finally, I document the economic conditions of the areas that received NMTCs.

Understanding the types of projects which secure investment and how those investments are financed helps contextualize the ongoing debate around place-based policies and targeted economic development more generally. The theoretical basis for place-based policies generally require the presence of market failures or the existence of externalities. Proponents of place-based policies have naturally looked to spatial deviations to justify their use. Without these deviations, then in markets with perfect labor mobility and inelastic housing supply the benefits of a location-specific subsidy or investment accrue to existing landowners, rather than residents (Neumark and Simpson 2015). With imperfect labor mobility, residents can potentially benefit but those benefits must be weighed against inefficient spatial investment, deadweight loss due to taxation, alternative use of public funds, and other costs. Urban economists have highlighted agglomeration economies, roughly defined as externalities arising due to the effect of population density on productivity, as a potential justification for place-based policies. Economists have identified and critiqued potential mechanisms underlying agglomeration economies such as thicker labor markets (Moretti 2011), knowledge spillovers (Moretti 2012), and industrial clusters (Duranton 2011). An alternative market failure centers on the spatial mismatch hypothesis which argues frictions in labor or housing markets limit resident mobility and their ability to take advantage of economic opportunities in other locations. This argument has historically been especially salient in the United States due to the country’s history of racism and red lining but similar issues have arisen in other countries due to race, immigration status, or caste. Critics of place-based policies, such as Glaeser and Gottlieb (2008), have raised concerns about the aggregate welfare implications of place-based policies, especially given the difficulty in understanding how elasticities of productivities vary spatially. By documenting the types of projects, the changing share of projects in urban areas, and sources and incentives for financing a particular place-based policy, I contribute towards an understanding of where market failures must exist to justify the continued use of place-based policies and the opportunity cost of that investment.

Studying the NMTC specifically, Gurley-Calvez et al. (2009) use investor behavior through 2004 to infer whether NMTC investment represented additional investment in low-income communities. They found encouraging results regarding individual investors, noting that at least some portion of investment represented new financial capital for low-income communities. However, for corporate investors, who claim over 99% of NMTCs in recent data, they find no evidence corporations are increasing investment but note corporations could be shifting investment from high-income to low-income communities. While Gurley-Calvez et al. (2009) examine changes in investment, Freedman (2012) examines the effect of the NMTC program on economic outcomes. He addresses the selection problem with a regression discontinuity design (RDD) that exploits the maximum income criteria for census tracts. He finds the NMTC reduces poverty and unemployment in low-income tracts. He also finds evidence the results may in part be due to changes in the composition of residents, thus highlighting the need for disaggregated data. Also utilizing an RDD, Harger and Ross (2016) examine the effect of the NMTC on business location and find evidence the NMTC increased the new businesses and new employment. They extend the analysis to examine effects by industry and, consistent with NMTC investment patterns, find increased employment and greater business creation in capital-intensive industries. Freedman and Kuhns (2018) focus on a specific industry of particular policy interest, retail food establishments, and find the NMTC program does induce growth in the retail food industry, especially among supermarkets. More recently, the Urban Institute recently released multiple detailed reports on the NMTC program which describing its effectiveness, how congressional funding has changed over time, unique data on types of projects, characteristics of the financial intermediaries which facilitate investment, and the program’s impact on local economic conditions and housing markets.^[1] I contribute to this literature by noting how some key characteristics of projects have changed over time, namely the share which are classified as real estate and the share of metropolitan projects. To the best of my knowledge this paper also provides the first discussion of NMTC claimants, their characteristics, and how they compare to taxpayers more generally.

In Section 2, I give a brief overview on the New Market Tax Credit program and describe how it has changed since it started. In Section 3, I describe the project and claimant data and provide summary statistics. In Section 4, I show how NMTC projects have changed from 2003 to 2019. In Section 5, I detail how NMTC claimants and their use of the credit have evolved since the program’s inception. In Section 6, to better understand how economic conditions correlate with tract selection, I use a logit model to quantify the marginal effect of changes in economic conditions on the observed probability a tract receives NMTC investment. Then, conditional on receiving investment, I examine to what extent economic conditions correlate with investment levels. Section 7 concludes.

2 The New Market Tax Credit

The New Market Tax Credit program was created by the Community Renewal Tax Relief Act of 2000. The program is administered by the Community Development Financial Institutions (CDFI) Fund at the U.S. Department of the Treasury. The CDFI Fund does not distribute credits per se but instead allocates, through a competitive process, investment authority to qualified applicants, called Community Development Entities (CDE), up to a congressionally approved limit. Investors receive a credit against their federal tax liability in exchange for investing in a CDE. The CDE acts as a financial intermediary which directs capital to qualified businesses in low-income communities. Investors receive credits spread over 7 years which total 39% of their investment. Since the program’s inception, the CDFI Fund has allocated over $65 billion worth of investment authority. For a more detailed description of the program, see Marples and Lowry (2019).

Businesses are eligible to receive an investment by a CDE, often a loan with preferential terms, if they are located in a qualifying census tract. In general, a census tract qualifies if it has a poverty rate of at least 20%, has a median family income below 80% of the statewide family income, or, if in an MSA and has a median family income below 80% of the metropolitan area median income.^[2] Approximately 43% of the census tracts in the US are eligible to receive NMTC investments. Eligible tracts include approximately 40% of the U.S. population.

Figure 1 displays the value (in 2020 dollars) of investment authority granted by Congress since the program’s enactment. Since 2006 the CDFI has had an annual investment authority for NMTC investments of at least $3.5 billion nominally. The Gulf Opportunity Zone Act of 2005 increased the nominal investment authority to $5 billion with a share of the additional funding directed to the gulf coast regions impacted by hurricanes Katrina, Rita, and Wilma. The Consolidated Appropriations Act of 2021, signed by President Trump in December of 2020, increased the annual investment authority to $5 billion through 2025. President Biden has proposed making the program permanent, with the annual amount indexed to inflation after 2025. The NMTC program has traditionally received bipartisan support, with credits authorized in every year since the program’s inception.

Figure 1:

Investment authority. The figures above displays the investment authority in 2020 dollars granted by Congress. Investment authority is allocated by the Community Development Financial Institutions (CDFI) Fund at the U.S. Treasury Department to Community Development Entities (CDE). As discussed in Section 2, each dollar of investment authority generates a potential 39 cents of New Market Tax Credits for investors in CDEs.

3 Data

In order to characterize NMTC projects and claimants I rely on two primary data sources. NMTC project data comes from the CDFI Fund at the U.S. Department of the Treasury. The CDFI data covers all NMTC projects through 2019. The CDFI data is at the project level and includes the census tract, investment amount, whether the project is a real estate investment, and whether or not the census tract is in a county designated as metropolitan by the U.S. Office of Management and Budget. I exclude projects from U.S. territories. I have a final set of 6807 projects which span all 50 states and DC. Table 1 provides summary statistics for the project data. Total investment is the sum of Qualified Low-Income Community Investment (QLICI). Investment per eligible population is the total investment divided by the total population in census tracts where investment was made. Metropolitan is the percentage of investment in metropolitan areas. Real Estate is the percentage of investment in projects classified as Real Estate by the CDFI. Through 2019, CDEs reported approximately $56 billion in allocated investment, with 81% of investment occurring in metropolitan areas and 46% going towards real estate projects.

For claimant data, I utilize firm-level corporate tax data for the years 2003–2018 from the IRS’ Statistics of Income, a stratified random sample conducted annually on initial tax returns.^[3] Table 2 compares Corporations with NMTC claimants. Panel A includes average characteristics such as Gross Receipts, Taxable Income, Taxes Paid, Total Assets, and NMTC Claims.^[4] All values are reported in 2020 dollars. NMTC claimants have historically had substantially larger gross receipts, income, and owed more in taxes. In addition, NMTC claimants report higher total assets, approximately $148 billion compared to approximately $50 million for the average corporation. Panel B displays the share of income and share of firms by the income distribution. For the population of corporations, 83% of firms have income less than $1 million and 60% of income is reported by firms with income between $1 billion and $100 billion. For NMTC claimants, the distribution of income is even more greatly skewed, with approximately 98% of income reported by firms with over $1 billion in annual income. Panel C displays the share of income and share of firms distributed across industries. NMTC claimants are concentrated in Financial services, with 72% of NMTC claimant income and 83% of NMTC-claiming firms being classified as Financial.

4 Project Characteristics

The selection and location of place-based policies provides both an appeal to policy makers and a challenge to researchers. Understanding the characteristics of past NMTC investments and how those characteristics have changed over time is valuable for evaluating and researching the effectiveness of the program. The NMTC program has evolved considerably since its inception and often in response to specific criticisms about the types of investment and location of the investment.

NMTC investments have and continue to be widely distributed geographically. Between 2003 and 2019, the NMTC authorized investments for over 6800 projects which span all 50 states and DC. Figure 2 displays investment per eligible population in tracts where investments have been made.^[5] States vary substantially in how much NMTC investment per eligible population they have received. Figure 3 displays how the total investment in different geographic regions have changed over time.^[6] While regional investment has been fairly steady, the South, and to a lesser extent the Midwest, have received a growing share of investment while the share of NMTC investment in the Northeast has declined.

Figure 2:

Investment per eligible population. The figure above displays investment per eligible population. Investment per eligible population is defined as the total Qualified Low-Income Community Investment (QLICI) investment in the state divided by the total population in census tracts where investments were made.

Figure 3:

Regional shares of investment. The figure above illustrates the regional share of NMTC investment for 2003 to 2019 using data from the Community Development Financial Institutions (CDFI) Fund. NMTC investment is defined as Qualified Low-Income Community Investment as reported by CDFI. Regions are defined according to the U.S. Census Bureau definitions.

Early criticism of the NMTC program focused on the large share of investment in metropolitan areas. In the Tax Relief and Health Care Act of 2006 (P.L. 109–432), Congress required the NMTC be allocated so non-metropolitan areas receive a proportional amount of investment allocation. Figure 4 displays the share of investment (QLICI) in metropolitan areas. Consistent with the regulatory requirement, by 2008 CDFI had substantially lowered the metropolitan share. Today the metropolitan share of investment is less than 75% of total investment, a decline of over 20 percentage points from the early years of the program. Figure 5 displays the share of investment in Real Estate since the program’s inception. Although not specified by the 2006 act, concurrent with the fall in metropolitan investment, the share of investment in real estate projects has declined. During the early years of the program, nearly 70% of investment was made in real estate projects but by 2019, investment in real estate projects had declined to approximately 32%. Projects often utilize NMTC investment to obtain more advantageous real estate financing and even projects classified as real estate may serve communities in a variety of ways. For a more complete description of NMTC project type and city-level data see Abravanel et al. (2013), Theodos et al. (2021b), and Theodos et al. (2021a).

Figure 4:

Metro Areas Share of Investment. The figure above illustrates the metropolitan share of NMTC investment for 2003 to 2019 using data from the Community Development Financial Institutions (CDFI) Fund. NMTC investment is defined as Qualified Low-Income Community Investment. Census tracts are classified as metropolitan if they are part of a metropolitan county, as defined by the Office of Management and Budget.

Figure 5:

Real estate share of investment. The figure above illustrates the real estate share of NMTC investment for 2003 to 2019 using data from the Community Development Financial Institutions (CDFI) Fund. NMTC investment is defined as Qualified Low-Income Community Investment. CDFI reports investments as either real estate, non-real estate, or investments in other CDEs.

5 Claimant Characteristics

The structure of the NMTC program provides incentives for investment in low-income census tracts. Through their investment in CDEs, NMTC claimants provide the financial capital the program attempts to redirect towards impoverished places. It is worth examining who claims the tax credits for at least two reasons. First, a primary concern about place-based programs is that the financial capital they attract is simply redirected from other investments. Although this paper makes no claims about where NMTC investments would have gone otherwise, a necessary step to understanding the potential trade offs is understanding the source of the financial capital. Second, credits such as the NMTC are known as General Business Credits (GBCs). A firm only uses a GBC they qualify for when they have positive tax liability. Moreover, GBCs can be carried forward up to 20 years to offset future tax liability. Other GBCs include credits related to research and development, renewable energy, certain oil and natural gas activities, low-income housing, and many other activities. Census tracts eligible for NMTC investment are often eligible for other place-based policies. It is beyond the scope of this paper to comprehensively examine interactions between GBCs but Konda et al. (2020) offer a broader discussion of GBCs and how their outstanding stock impacts utilization. When a firm chooses to qualify and utilize NMTCs they are inherently trading off between other investment opportunities and other credits potentially available to them. To better understand the relative value of different GBCs, it is valuable to understand the characteristics of firms utilizing the NMTC.

Figures 6 and 7 expand on the characteristics highlighted in Table 2 by displaying how claimant’s reported assets and income have changed since the program’s inception. All figures are shown in inflation-adjusted terms using 2020 dollars. Assets and income of claimants follow a similar pattern throughout the period, with each experiencing substantial increases since 2009. In 2018, NMTC claimants reported average assets of approximately $244 billion and over $15 billion in total income.

Figure 6:

NMTC claimants assets. The figure above graphs the average amount of total assets reported on Form 1120 for the years 2003–2018 by corporations who claim the New Market Tax Credit on Form 3800. Values are calculated using corporate tax data from the IRS Statistics of Income files, as described in Section 3.

Figure 7:

NMTC claimants income. The figure above graphs the average amount of total income reported on Form 1120 for the years 2003–2018 by corporations who claim the New Market Tax Credit on Form 3800. Values are calculated using corporate tax data from the IRS Statistics of Income files, as described in Section 3.

The NMTC has consistently been a greater percentage of income for smaller claimants. Figure 8 graphs the total share of taxable income (solid line) as well as the percentage of taxable income for small (dotted line) and large (dashed line) claimants. A firm is classified as a large claimant if it has assets greater than the median-sized claimant while small claimants are those with assets less than the median-sized claimant. Consistent with their volume of credits, large claimants’ largely track with total share, whereas NMTCs are a larger share of taxable income for small claimants throughout the sample period.

Figure 8:

NMTC Pct. of taxable income. The figure above displays the New Market Tax Credit share of taxable income for NMTC claimants overall (solid line) and broken into large and small claimants. Large claimants (dashed line) are those with assets greater than the median. Small claimants (dotted line) are those with assets less than the median. Values are calculated using corporate tax data from the IRS Statistics of Income files, as described in Section 3.

Figure 9 displays the NMTC share of total reported GBCs for NMTC claimants. Initially NMTCs served as a small fraction of the GBCs reported by claimants and were less than 6% of claimant’s total GBCs. By 2019, the NMTC had grown to approximately 45% of GBCs for NMTC claimants. The NMTC increased share of GBCs for claimants may be due to either a change in the relative value of the credit or because the composition of claimants has changed.

Figure 9:

Pct. of GBCs for NMTC claimants. The figure above displays the share of General Business Credits that were New Market Tax Credits for NMTC claimants. Values are calculated using corporate tax data from the IRS Statistics of Income files, as described in Section 3.

Financial firms have historically claimed the vast majority of credits. Figure 10 graphs the share of New Market Tax Credits claimed by financial firms since 2003. Financial firms’ claimed a smaller share of NMTCs during and after the 2007–2008 recession as their tax liability declined. Since then, and consistent with their ability to carry credits forward, financial firms have claimed a higher percentage of NMTCs.

Figure 10:

Pct. of NMTCs for financial firms. The figure above displays the share of New Market Tax Credits which were claimed by financial firms. Values are calculated using corporate tax data from the IRS Statistics of Income files, as described in Section 3.

6 Regression Analysis

7 Conclusion

The NMTC has changed considerably since the program began both in terms of investments and how the credit is used. Projects have shifted to be less metropolitan and are less likely to be classified as Real Estate. Geographically, NMTC investments continue to be made in all 50 states with a slightly larger share of investments in the South and Midwest. Corporations with assets and incomes substantially higher than the average corporation continue to be most likely to claim the credit. Partially a reflection of broader macroeconomic improvements, average assets and income for claimants has grown considerably since 2009.

The cross-sectional state fixed effect regressions in Section 6.1 provide evidence worse economic conditions, and especially a worse poverty rate, lead to a higher probability of receiving NMTC investment. This is true even after controlling for satisfying the threshold conditions necessary for qualification. The correlation however is not present in regressions utilizing variation across time. More generally, it is less clear how tightly correlated economic conditions are with the amount of NMTC investment a tract receives. The results should help put to rest the concern projects are overly concentrated in urban or metropolitan areas.

Place-based policies remain a popular policy worthy of further study. This paper focused on correlations of characteristics explicitly considered by CDFI. These can help inform modeling choices on the selection process and the levels of observed investment. This paper has not addressed factors which are unobservable to CDFI but may be observed by program participants. Understanding the extent to which investments are targeted according to these unobservables is an important next step for research on both program design and understanding the effectiveness of place-based policies. Substantial work remains on credibly identifying effects on residents and non-residents, spillovers to non-selected geographies, and disentangling underlying mechanisms. Progress on any of these topics would be valuable, especially given the continued and widespread use of place-based policies.