This is the accessible text file for GAO report number GAO-07-296 entitled 'Tax Policy: New Markets Tax Credit Appears to Increase Investment by Investors in Low-Income Communities, but Opportunities Exist to Better Monitor Compliance' which was released on January 31, 2007. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. Report to Congressional Committees: United States Government Accountability Office: GAO: January 2007: Tax Policy: New Markets Tax Credit Appears to Increase Investment by Investors in Low-Income Communities, but Opportunities Exist to Better Monitor Compliance: GAO-07-296: GAO Highlights: Highlights of GAO-07-296, a report to congressional committees Why GAO Did This Study: The Community Renewal Tax Relief Act of 2000 authorized up to $15 billion of allocation authority under the New Markets Tax Credit (NMTC) to stimulate investment in low-income communities. The act mandated that GAO report on the program to Congress by January 31, 2004, 2007, and 2010. Two subsequent laws authorized an additional $1 billion in NMTC authority for certain qualified investments and extended the program for 1 year with an additional $3.5 billion of authority. This report (1) describes the status of the NMTC program, (2) profiles NMTC program participants, (3) assesses the credit’s effectiveness in attracting investment by participating investors, and (4) assesses IRS and the Community Development Financial Institutions (CDFI) Fund compliance monitoring efforts. To conduct the analysis, GAO surveyed NMTC investors, conducted statistical analysis, and interviewed IRS and CDFI Fund officials. What GAO Found: As of January 2007, the CDFI Fund had awarded $12.1 billion of NMTC authority to 179 Community Development Entities (CDE). CDEs that received allocations began making NMTC investments in 2003, and the program has continued to grow since then. Investors use two main investment structures to make NMTC investments: direct investments to CDEs and tiered investments, which include equity investments and leveraged investments, where a portion of the investment amount originates from debt and a portion from equity. Banks and individuals constitute the largest proportion of NMTC investors, though banks and other corporations have made the largest share of NMTC investment. CDEs that received allocations applied for allocations in a competitive selection process and, through fiscal year 2005, most investment from CDEs to low-income communities had been used for either commercial real estate rehabilitation or new commercial real estate construction. Figure: NMTC Loans and Investment by Type of Activity for Fiscal Years 2003 through 2005: [See PDF for Image] Source: GAO analysis of CDFI Fund data. [End of Figure] The results of GAO’s survey and statistical analysis indicate that the NMTC may be increasing investment in low-income communities by participating investors. Investors indicated that they have increased their investment budgets in low-income communities as a result of the credit, and GAO’s analysis indicates that businesses may be shifting investment funds from other types of assets to invest in the NMTC, while individual investors may be using at least some new funds to invest in the NMTC. The CDFI Fund and IRS developed processes to monitor CDEs’ compliance with their allocation agreements and the tax code. However, IRS’s study of CDE compliance does not cover the full range of NMTC transactions, focusing instead on transactions that were readily available, and may not support the best decisions about enforcement in the future. Moreover, IRS and the CDFI Fund are not collecting data that would allow IRS to identify credit claimants and amounts to be claimed. What GAO Recommends: To ensure that it is reviewing the full range of NMTC transactions, IRS should develop information for selecting which CDEs to audit as part of its compliance study. In addition, IRS should work with the CDFI Fund to further explore options for cost effectively monitoring investor compliance. IRS and the CDFI Fund agreed with our recommendations. [Hyperlink, http://www.gao.gov/cgi-bin/getrpt?GAO-07-296]. To view the full product, including the scope and methodology, click on the link above. For more information, contact Michael Brostek at (202) 512-9110 or brostekm@gao.gov. [End of Section] Contents: Letter: Results in Brief: Background: CDEs Are Using NMTC Allocations to Invest in Low-Income Communities, and the CDFI Fund Is Tracking Program Implementation: Financial Institutions and Individuals Are the Primary NMTC Investors, and CDEs Most Often Use NMTC Investments to Make Loans to Qualified Businesses: NMTC Investors Report That the NMTC Increases Investment in Low-Income Communities and Statistical Analysis Indicates That These Investments May Be Financed by Shifting Assets from Other Uses and Some New Investment: IRS and the CDFI Fund Monitor NMTC Compliance, but Additional Opportunities Exist to Better Measure Noncompliance and Identify NMTC Investors: Conclusions: Recommendations for Executive Action: Agency Comments: Appendix I: Objectives, Scope, and Methodology: Appendix II: Description of Data and Methodology for Statistical Analysis of the Effect of NMTC Participation on Investment: Appendix III: NMTC Investment Data by State, Fiscal Years 2003 through 2005: Appendix IV: Comments from the Community Development Financial Institutions Fund: Appendix V: Comments from the Internal Revenue Service: Appendix VI: GAO Contact and Staff Acknowledgments: Tables: Table 1: NMTC Allocation Rounds: Table 2: NMTC Claimant Types: Table 3: Reasons NMTC Investors Invested in the NMTC Program (in Percentages): Table 4: Investor Knowledge of CDE Operations (in Percentages): Table 5: NMTC Allocations Awarded by Round: Table 6: CDEs That Applied for NMTC and Received Allocations by Round: Table 7: Top 10 States by NMTC Dollars through Fiscal Year 2005: Table 8: Effects of NMTC Individual Investor Participation on Wealth: Table 9: Growth in Net Assets Using Fixed Effects Regression and Comparisons Based on Nearest Neighbor Propensity Score Matching: Table 10: Baseline Analysis: Instrumental Variables Fixed Effects Regressions on the Full Sample: Table 11: Growth in Assets: Comparisons Based on Nearest Neighbor Propensity Score Matching: Figures: Figure 1: NMTC Process for Using Allocated Tax Credits to Make Qualified Low-Income Community Investments: Figure 2: NMTC Eligible Areas: Figure 3: Number of CDE Allocations by Round (Calendar Year): Figure 4: Qualified Equity Investment by Calendar Year: Figure 5: Comparison of NMTC Investment Structures: Figure 6: Interaction of CDFI Fund NMTC Data Collection Systems: Figure 7: NMTC Loans and Investment by Type of Activity for Fiscal Years 2003 through 2005: Figure 8: NMTC Dollars Used in Loans with Better Rates and Terms: Figure 9: NMTC Investors Packaging the NMTC with Other Government Incentives: Figure 10: Activities Investor Survey Respondents Undertake to Monitor CDE Compliance: Abbreviations: AAS: Allocation Agreement System: ATS: Allocation Tracking System: BRTF: Business Returns Transaction File: CDE: Community Development Entity: CDFI: Community Development Financial Institutions: CIIS: Community Investment Impact System: CPI: Consumer Price Index: CRA: Community Reinvestment Act: FCOS: Financial Counseling and Other Services: FEMA: Federal Emergency Management Agency: GO Zone: Gulf Opportunity Zone: HUB Zone: Historically Underutilized Business Zone: ILR: Institution Level Report: IRS: Internal Revenue Service: IRTF: Individual Returns Transaction File: MOU: memorandum of understanding: NCMS: New Markets Compliance Monitoring System: NMTC: New Markets Tax Credit: SBA: Small Business Administration: SCF: Survey of Consumer Finances: TIN: Taxpayer Identification Number: TLR: Transaction Level Report: QALICB: qualified active low-income community business: QEI: qualified equity investment: QLICI: qualified low-income community investment: United States Government Accountability Office: Washington, DC 20548: January 31, 2007: Congressional Committees: Congress established the New Markets Tax Credit (NMTC) program in the Community Renewal Tax Relief Act of 2000[Footnote 1] as part of an ongoing effort to address one of our nation's most persistent challenges--the revitalization of impoverished, low-income communities. Conventional access to credit and investment capital for developing small businesses, retaining jobs, and revitalizing neighborhoods is often limited in economically distressed communities or in communities with large low-income populations. The NMTC provides investors (individuals, financial institutions, other corporations, etc.) with a tax credit for investing in communities that are economically distressed or consist of low-income populations. Currently, the Community Development Financial Institutions (CDFI) Fund in the Department of the Treasury is authorized to allocate up to $19.5 billion[Footnote 2] in tax credit authority to Community Development Entities (CDE) that manage NMTC investments in low-income community development projects. CDEs are domestic corporations or partnerships with a primary mission of serving or providing investment capital for low-income communities or low-income persons. Tax credit authority is the amount of investment for which investors can claim a tax credit at rates that total, over the 7 years they can claim the credit, 39 percent of their investment. In return for the tax credit, investors supply capital to the CDEs, which, in turn, make investments in qualified low-income communities. The Community Renewal Tax Relief Act of 2000 mandated that we report to Congress on the NMTC program by January 31, 2004, 2007, and 2010. In our report issued January 30, 2004,[Footnote 3] we described the status of the NMTC program, profiled CDEs that received first round allocations (there have now been four rounds of NMTC allocations), and evaluated whether the systems were in place or planned in order to ensure NMTC compliance. We concluded progress was being made in implementing the NMTC program, but we also recommended that Internal Revenue Service (IRS) and the CDFI Fund work together to develop plans for designing and implementing compliance monitoring processes. IRS and the CDFI Fund agreed with our recommendation and have taken steps to design and implement compliance monitoring processes. Based on consultations with staff at cognizant congressional committees, this report (1) describes the status of the NMTC program; (2) profiles the characteristics of NMTC investors, the CDEs that receive NMTC allocations, and the businesses and communities that receive NMTC investments; (3) assesses how effective the NMTC has been in bringing new investment to low-income communities by the investors that have participated in the program; and (4) assesses the steps that IRS and the CDFI Fund are taking to ensure CDEs and investors are complying with the NMTC and evaluates how effective these steps have been. To accomplish these reporting objectives, we met with officials from the CDFI Fund and IRS. We collected documents on the program's status and efforts to monitor NMTC compliance. We also analyzed data from the CDFI Fund on the CDEs and their investment in low-income communities and tax return data from tax years 1997 through 2004 for investors in the NMTC program. We used these data to report summary statistics that profile the participants in the program and to conduct statistical analysis that measures the effect of the NMTC on investment by participating investors. In our statistical analysis, we compared a stratified random sample of taxpayers that did not make NMTC investments with investors that did make NMTC investments using fixed- effects regressions and comparisons based on other statistical methods to measure the effect of the NMTC on corporate investors' growth in net assets and individual investors' growth in wealth. We also surveyed investors in the NMTC program in order to provide additional information on the effect of the credit and characteristics of the investors. Our overall response rate was 51 percent. We weighted our survey responses using information on investor type and investor size to reduce possible nonresponse bias that is associated with investor type and size. Results from our statistical analysis and the survey are limited to the effects of NMTC investments on the investment choices of participating investors and do not assess the effect of this investment on the investments by non-NMTC participants in low-income communities. Our scope and methodology section (app. I) provides additional details on how we did our work. Our work was conducted from July 2006 through December 2006 in accordance with generally accepted government auditing standards. In December 2006, we requested written comments on a draft of this report from the Director of the Community Development Financial Institutions Fund and the Commissioner of the Internal Revenue Service; their comments are reprinted in appendices IV and V. Results in Brief: Since the CDFI Fund made its first allocations in 2003, the NMTC program has grown in terms of the amount of tax credit authority allocated to CDEs, the complexity of NMTC investments, and the amount of money invested in low-income communities. As of January 2007, the CDFI Fund had made 233 NMTC allocation awards totaling $12.1 billion in allocation authority to 179 CDEs--some CDEs have received multiple allocations--which the CDEs have used to attract nearly $5.3 billion in NMTC investment. These CDEs with allocation awards are required to attract investment sufficient to use the remaining $6.8 billion of allocation authority in the coming years. The total amount per year invested by these CDEs in low-income communities grew from about $140 million in 2003 to $2.2 billion in 2005. As the NMTC program has grown, more investors have participated in more complicated NMTC investment structures, such as tiered investments, which include both equity investments and leveraged investments. The CDFI Fund has developed data systems that track allocation agreements (which set forth conditions such as approved uses of the allocations and approved service areas), allocated credits, and collected data about investors, the CDEs, and their investments in low-income communities. The CDFI Fund combines data from these systems to monitor compliance with allocation agreements and to help IRS determine whether laws and regulations are being observed. All of these systems were operational in time to meet the CDFI Fund's needs. Banks and individuals constitute the majority of NMTC claimants, accounting for 70 percent of NMTC claimants through 2006, though banks and other corporations account for the largest share of NMTC investment. Banks and other corporations that invested in the credit had relatively large net assets, and individuals who invested in the NMTC had, on average, higher incomes than other taxpayers. Most investors made only one investment in a CDE: 55 percent of investors made a single investment while 12 percent made five or more investments. The CDEs applied for far more allocation dollars than were available. They received only about 11 percent of $107 billion in allocation authority for which they applied. Data reported through fiscal year 2005 indicate that businesses in low-income communities received investments from CDEs to fund over 580 NMTC projects, totaling over $3 billion of investment. The projects were funded primarily by loans from the CDEs and were used chiefly to finance commercial real estate construction and rehabilitation. The communities where the investment projects were located were dispersed across states and about 90 percent were located in areas designated as "areas of high distress" because of factors such as low median incomes or high unemployment rates. The results of our survey and statistical analysis are consistent with the NMTC program increasing investment in eligible low-income communities by the investors that participate in the program and with this investment coming primarily from funds shifted from other uses. Such a shift would be one indicator that the NMTC program is effective because the NMTC sought to increase investment in eligible low-income communities. An estimated 64 percent of the NMTC investors reported that they increased the share of their investment budget for low-income communities because of the credit. One limitation of our survey is that the population of NMTC investors we surveyed benefit from claiming the credit and have an interest in ensuring that the NMTC program continues in the future. However, in many cases the survey also indicated that the credit alone may not have been sufficient to justify the investment and meeting other government regulations may be an important incentive for making NMTC investments. Any increased investment in low-income communities because of the credit can occur when NMTC investors make new investment by increasing their total funds available for investment or when they shift funds from other uses in higher income communities. Our statistical analysis suggests that in general corporate NMTC investors are not increasing their overall level of investment to participate in the NMTC program. Taking this information together with information from our survey of investors, we infer that the most likely effect of the credit is that corporate investors, which make the majority of investments in CDEs, are shifting investment into low- income communities from higher income communities. Our statistical analysis indicates that unlike corporate investors, participating individual investors as a group appear to be making at least some new investment to participate in the NMTC program. This finding that corporate and individual NMTC investors appear to be increasing investment in low-income communities is not, in and of itself, sufficient to determine that the credit is effective. For example, it was beyond the scope of our analysis to determine whether investment by NMTC investors reduced such investments by non-NMTC investors. A complete evaluation of the program's effectiveness also requires determining the costs of the program, including any behavioral changes by taxpayers that may be introduced by shifted investment funds. In addition, such an evaluation requires an assessment of the program's economic and social benefits. For example, to the extent a community experiences a reduction in poverty and increases in employment opportunities as a result of the program, possible "spillover" benefits to the community may include reductions in crime and improvements in the health status of community residents. The CDFI Fund is working with a contractor to develop plans for a comprehensive evaluation of the NMTC, which may include evaluating the program's effectiveness. IRS and the CDFI Fund have taken steps to monitor compliance with the requirements of the NMTC program, but additional opportunities exist to better measure noncompliance and identify NMTC investors. IRS is conducting a compliance study focusing on whether CDEs comply with the "substantially all" test imposed by the Internal Revenue Code, which requires that CDEs invest at least 85 percent of a qualified equity investment (QEI) in a low-income community within 1 year of receiving the investment. However, because CDEs did not file initial returns as soon as IRS expected, IRS was not able to select CDEs to audit in a way likely to produce findings that are representative of the full range of CDE activity. However, as the program expands and more CDEs make NMTC investments, IRS should have more CDEs to choose from when selecting CDEs to audit for its compliance study, and IRS could use CDFI Fund data to aid in developing criteria for selecting which CDEs to audit. The CDFI Fund is focusing on ensuring that CDEs fulfill their allocation agreement requirements. The CDFI Fund monitors CDE compliance primarily through its data systems and, to a lesser extent, by making site visits. The data systems are designed to enable the CDFI Fund to identify when a CDE falls out of compliance with its allocation agreement. However, neither IRS nor the CDFI Fund currently have sufficient information to enable the IRS to identify NMTC investors and the amount of credit that the investors are entitled to claim, particularly when the original investments are sold to others. CDEs may be a useful source of information because they need to know who their investors are, even when investments are sold, in order to submit appropriate reports to those investors. If IRS or the CDFI Fund developed ways to identify investors and the amounts they invested, even when NMTC investors sell their equity share in a CDE after the original investment is made, the IRS would be better able to ensure that credits are claimed correctly. To ensure that IRS is reviewing the full range of NMTC transactions and that the conclusions of its compliance study are more representative of all CDEs with NMTC allocations, we recommend that IRS use CDFI Fund data and the results of its current NMTC compliance study to develop criteria for selecting which CDEs to audit as part of its future compliance monitoring efforts. Additionally, to ensure that eligible taxpayers claim the correct amount of NMTC on their tax returns and IRS is able to identify all tax credit claimants in the event of a CDE falling out of compliance with NMTC regulations, we recommend that IRS work with the CDFI Fund to further explore options for cost effectively monitoring investor compliance and developing a way to identify NMTC claimants, even in instances where the original investor sells its equity share in a CDE, and the amount of NMTC investment that investors made. In commenting on this report, both the Acting Director of the CDFI Fund and the Commissioner of Internal Revenue agreed with our recommendations (their responses are reprinted in appendices IV and V). Background: As we noted in a past report, the NMTC was created in an effort to increase the amount of capital available to low-income communities,[Footnote 4] facilitate economic development in these communities, and encourage investment in high-risk areas.[Footnote 5] In order to achieve these goals, the program allows investors that provide eligible capital to low-income communities and businesses to reduce their tax liability by 39 percent of the amount of the investment over a 7-year period. The NMTC Investment Process: The process of making an NMTC investment involves several steps and a number of stakeholders. Before applying for an NMTC allocation, the applicant must apply for and be certified as a CDE, which is an entity that manages investments for community development.[Footnote 6] Once an organization has been certified as a CDE by the CDFI Fund, it is then eligible to apply for an NMTC allocation. Both for-profit and nonprofit CDEs may apply for and receive NMTC allocations (once a CDE is awarded with an allocation, it is often referred to as an allocatee). However, only a for-profit CDE can offer NMTCs to investors. Therefore, when a nonprofit CDE receives an NMTC allocation, it must transfer the allocation to one or more for-profit subsidiary CDEs (referred to as suballocatees). NMTC applicants submit standardized application packages in which they respond to a series of questions about their track records, the amounts of NMTC allocation authority being requested, and their plans for using the tax credit authority. The CDFI Fund staff and a group of external reviewers who have experience in business, real estate, and community development finance then review the applications and score them based on the following four areas: (1) community impact, (2) business strategy, (3) capitalization strategy, and (4) management capacity. The applicants can receive a score of up to 25 points in each of the areas, and CDEs can obtain up to 10 additional "priority points" for demonstrating that they have track records of successfully investing in low-income communities and/ or that they intend to invest in unrelated entities. After being reviewed and scored by three different reviewers (and, in some cases, a fourth reviewer if a scoring anomaly exists), the applicants are ranked and NMTC allocation awards are made in descending order of the highest aggregate scores to applicants that met minimum thresholds in each of the four areas.[Footnote 7] The CDFI Fund makes award determinations in this order until the allocation authority is exhausted. The CDFI Fund also provides a written debriefing to each CDE that does not receive an allocation in order to provide them with reasons their application did not receive an NMTC award and to provide the CDE with suggestions on how to be more competitive for NMTC awards when applying in future rounds. As figure 1 shows, after the allocations are made to the CDEs, investors make equity investments, by acquiring stock or a capital interest, in the CDEs to receive the right to claim tax credits on a portion of their investment.[Footnote 8] In turn, the CDE must invest "substantially all"[Footnote 9] of the proceeds into qualified low- income community investments (QLICI). Eligible investments include, but are not limited to, loans to or investments in businesses to be used for developing residential, commercial, industrial, and retail real estate projects; and purchasing loans from other CDEs. Figure 1: NMTC Process for Using Allocated Tax Credits to Make Qualified Low-Income Community Investments: [See PDF for image] Source: GAO. [A] Only a for-profit CDE can receive qualified equity investment from NMTC investors. These CDEs can then make investments in other CDEs that could be for-profit CDEs or nonprofit CDEs or they can directly invest the NMTC funds in low-income communities. However, both for-profit and nonprofit CDEs can receive allocations from the CDFI Fund. If a nonprofit CDE receives a NMTC allocation from the CDFI Fund, it must transfer the allocation authority to a for-profit CDE before NMTC investments can be made. [End of figure] Once a qualifying investment has been made in a CDE and the CDE has invested the funds in an eligible low-income community, the investor can claim the tax credit over the course of 7 years. In addition, equity investors may receive returns on their investments in the form of dividends or other income that they receive from the CDE during the period in which they are eligible to claim the credit. The NMTC investor is still usually allowed to claim the NMTC for the full 7-year period even if the business that the CDE provides investment to defaults on its loans or files for bankruptcy. However, in the case of a business that receives NMTC funds going bankrupt, the ability of the investor to recover its initial equity investment in a CDE would depend on the assets and financial condition of the CDE as well as the original agreement that the CDE entered into with the investor. The NMTC is a nonrefundable tax credit, meaning that taxpayers do not receive payments for tax credits that exceed their total tax liability. In addition, taxpayers that are eligible to claim the tax credit may sell their investment, along with the right to claim any remaining tax credits, to another investor after the initial NMTC investment. For example, an investor may make an equity investment in a CDE that would allow it to claim the credit and then sell its equity share in the CDE to another investor, thereby transferring the right to claim the remaining credits to this investor. The original investor may choose to sell its equity share in a CDE, and consequently its right to claim the credit, because it does not have a tax liability for that year or other reasons, such as the timing of the original investment.[Footnote 10] Once investors begin claiming the credit on their tax returns, three things can trigger a recapture event (meaning that the investor will no longer be able to claim the credit because the investment no longer qualifies for NMTCs). The NMTCs can be subject to a recapture if the CDE (1) ceases to be certified as a CDE, (2) does not satisfy the "substantially all" requirement, or (3) redeems the investment. In general, a recapture event means that the investors that originally purchased the equity investment and subsequent holders of the investment are required to increase their income tax liability by the credits previously claimed plus interest for each resulting underpayment of tax. Legislative Changes Created Targeted Populations: Two recent legislative changes have increased the number of areas where NMTC investments can be made. First, the American Jobs Creation Act of 2004[Footnote 11] added "targeted populations" to the eligibility criteria for NMTC investments. Second, Congress also expanded the NMTC program in 2005,[Footnote 12] providing an additional $1 billion of allocation authority to be made available to CDEs with a significant mission of recovery and redevelopment of low-income communities in the Gulf Opportunity Zone (GO Zone), which are specified areas in Louisiana, Mississippi, and Alabama that were affected by Hurricane Katrina during 2005. In general, targeted populations were introduced to give CDEs flexibility in making investments serving individuals and groups that reside or work in communities that might not otherwise fall under the NMTC program's geographically based definition of a low-income community. Currently, regulations defining targeted populations have not been finalized. However, the CDFI Fund and IRS have provided guidance for what qualifies as a targeted population.[Footnote 13] These guidelines specify that the targeted populations, which are individuals or an identifiable group of individuals, must meet tests to qualify as low-income communities and the businesses or entities receiving the investments must also meet certain criteria.[Footnote 14] In IRS's recently provided guidance, the definition of GO Zone targeted populations is similar to the definition for low-income targeted populations with some differences. In cases where a business is located within the GO Zone, it does not mean that it automatically qualifies for NMTC investment dollars. First, the GO Zone targeted population need not qualify as low-income individuals as defined above, but rather the population must consist of individuals who lack access to loans or equity investments because they were displaced from their principal residence or lost their principal source of employment because of Hurricane Katrina. Second, the NMTC investment must serve targeted populations in census tracts within the GO Zone that meet certain requirements, including that they contain one or more areas designated by the Federal Emergency Management Agency (FEMA) as flooded or having sustained extensive or catastrophic damage as a result of Hurricane Katrina. Figure 2 illustrates the effect that recent legislative changes have had on the census tracts that are eligible to receive NMTC investments. As the figure shows, geographically, a large portion of the country qualifies for NMTC investment, and there are eligible areas in every state. The figure also shows the area of the GO Zone where NMTC investments can be made in both eligible low-income communities and specified targeted populations as a result of additional allocation authority made available for areas affected by Hurricane Katrina. Figure 2: NMTC Eligible Areas: [See PDF for image] Source: CDFI Fund. Note: All unshaded areas identified as "Not NMTC eligible" could receive NMTC investment funds if CDEs serve targeted populations in those areas under the American Jobs Creation Act of 2004 (Pub. L. No. 108-357). In addition, targeted populations in areas shaded in black in the GO Zone may receive NMTC investment because they meet the definition of a GO Zone targeted population. [End of figure] Authorized Allocation Rounds End in 2008: Congress initially provided a schedule for allocating annual NMTC authority to CDEs for calendar years 2001 through 2007.[Footnote 15] However, as we also reported in 2004, the CDFI Fund did not make any NMTC allocations to CDEs until 2003 because it needed to complete various start-up tasks for the new program, such as establishing the rules for using allocations. Because the initial allocations were not made until 2003, the CDFI Fund combined the allocation amounts available for 2001 and 2002 and awarded those NMTC allocations in 2003. The allocation amounts designated for 2003 and 2004 were then combined and awarded in 2004. Table 1 shows the current schedule for allocation rounds. Since 2004, allocation awards have been made to CDEs annually. Table 1: NMTC Allocation Rounds: Dollars in billions. Round: Round 1; Allocation year: 2003; Original allocation years: 2001-2002; Amount allocated: $2.5. Round: Round 2; Allocation year: 2004; Original allocation years: 2003-2004; Amount allocated: 3.5. Round: Round 3; Allocation year: 2005; Original allocation years: 2005; Amount allocated: 2.0. Round: Round 4; Allocation year: 2006; Original allocation years: 2006; Amount allocated: 4.1[A]. Round: Round 5; Allocation year: 2007; Original allocation years: 2007; Amount allocated: 3.9[A]. Round: Round 6; Allocation year: 2008; Original allocation years: 2008; Amount allocated: 3.5[B]. Round: Total; Allocation year: $19.5. Source: CDFI Fund. [A] The amounts available to be allocated in Round 4 and Round 5 were increased by $600 million and $400 million respectively because of increased NMTC allocation limits targeted toward the GO Zone. [B] Congress initially only authorized NMTC allocation authority through 2007. However, the Tax Relief and Health Care Act of 2006 (Pub. L. No. 109-432) extended NMTC allocation authority for 1 year (through 2008) with an additional $3.5 billion of NMTC allocation authority. [End of table] As of January 2007, there have been four completed rounds of NMTC allocations, and the CDFI Fund is receiving applications for the 2007 round of NMTC allocation awards, which will be announced in September 2007. The 2007 allocation awards were originally scheduled to be the last authorized round of NMTC allocation awards. However, in December 2006, Congress passed and the President signed the Tax Relief and Health Care Act of 2006,[Footnote 16] which extends the NMTC for an additional year (through the end of 2008) with an additional $3.5 billion of NMTC allocation authority. Regulations are also required to be drafted to ensure that nonmetropolitan areas receive a proportional allocation of qualified equity investments. CDEs Are Using NMTC Allocations to Invest in Low-Income Communities, and the CDFI Fund Is Tracking Program Implementation: The CDFI Fund has completed four rounds of NMTC allocations, which CDEs are using to attract investment. The investment structures used to complete these deals have taken a variety of forms, including combining debt and equity in limited liability partnerships in order to invest in a CDE--called leveraging. In addition, the CDFI Fund has developed four main data collection systems to track efforts to implement and monitor the expanding NMTC program. NMTC Allocations and Investments in CDEs and Low-Income Communities Have Increased in Number and Amount: Beginning in 2003, the CDFI Fund awarded NMTC allocations of varying amounts to a number of CDEs. The CDFI Fund has awarded 233 NMTC allocations to 179 different CDEs totaling $12.1 billion over the course of the four completed NMTC allocation rounds. As figure 3 shows, the CDFI Fund made awards to the largest number of CDEs in 2003, when the fund awarded NMTC allocations to 66 CDEs, and it made awards to the smallest number of CDEs in 2005 when 41 CDEs received allocations. In its most recent allocation round in 2006, the CDFI Fund made allocations to 63 CDEs for a total of $4.1 billion of tax credit authority. The largest award to a single CDE in this allocation round was $143 million, while the median award was $60 million. Figure 3: Number of CDE Allocations by Round (Calendar Year): [See PDF for image] Source: GAO analysis of CDFI Fund data. [End of figure] The CDEs receiving allocations were able to attract an increasing number of QEIs. As of December 2006, investors had made nearly 1,400 QEIs in CDEs, and as more allocation rounds have taken place, the number of QEIs has grown. Relatively few QEIs were made in 2003 when the program was in its early stages, but the number of QEIs increased significantly in both 2004 and 2005. This pattern of growth reflects increases in NMTC allocation authority and increased time for CDEs to establish business relationships with potential investors. In addition, more QEIs were made in CDEs that received allocations in 2003 and 2004 than in CDEs that received NMTC allocations in 2005. As of December 2006, 749 QEIs had been made in first round NMTC allocatees, 478 QEIs had been made in second round NMTC allocatees, and 154 QEIs had been made in third round allocatees.[Footnote 17] As figure 4 shows, the CDEs were generally able to attract increasing dollar amounts of qualified equity investment. QEI grew from about $140 million of investment in 2003 to over $2.2 billion of investment in 2005, and as of mid-December 2006, CDEs had recorded nearly $1.5 billion in NMTC investment for the year--totaling $5.3 billion over the period. CDEs are required to invest the remaining $6.8 billion of allocation authority awarded to this point during the coming years. At the same time, the size of the QEIs varied considerably across CDEs. According to CDFI Fund data, the largest QEI made through December 2006 was $113 million, while the median QEI during this period was about $1.8 million. Figure 4: Qualified Equity Investment by Calendar Year: [See PDF for image] Source: GAO analysis of CDFI Fund data. Note: Amount of QEI in 2006 is through mid-December. [End of figure] The CDEs used this QEI to make investments in 583 qualified NMTC projects totaling $3.1 billion through fiscal year 2005.[Footnote 18] Nearly all of these investments have been to qualified active low- income community businesses (QALICBs) in qualifying areas. However, according to CDFI Fund data, a small number (about 1 percent) of the investments were made to other CDEs, as permitted under NMTC regulations. As more NMTC allocation awards are made and more NMTC investment transactions are completed, additional information will be available about the size and type of NMTC investments. The Ability of Investors to Use Tiered Investment Structures May Have Contributed to the Growth of the NMTC Program: Certain NMTC investment structures may have been a factor in the growth of the program by making NMTC investments more attractive. NMTC investors have used two primary investment structures when making QEIs in CDEs: (1) direct NMTC investment and (2) tiered NMTC investments.[Footnote 19] As of December 2006, about 54 percent of the $5.3 billion in NMTC investments were made using tiered investment structures. In a direct NMTC investment, an investor makes a QEI in a CDE that reinvests the money in a low-income community. (See fig. 5 for a description of these NMTC investment structures). In tiered investment structures, which include both equity investments and leveraged NMTC investments, investors provide equity or loans to a pass- through entity that combines funds from several sources, and the pass- through entity makes the QEI in a CDE.[Footnote 20] In both direct and tiered investment structures, equity investors in a CDE are able to claim the NMTC on their tax returns and, after leaving the equity investment in the CDE for the 7 years during which they are eligible to claim the credit, they can redeem their original equity stake in the CDE. In a tiered equity investment structure, the dollars invested in the investment fund consist entirely of equity investments from multiple investors. These investment structures accounted for about 13 percent of NMTC investment as of December 2006. In a tiered leveraged investment structure,[Footnote 21] a portion of the money being invested in the investment fund comes from equity investors and a portion of the money originates from a debt investment (loan). As of December 2006, about 41 percent of all NMTC investment was made using the leveraged approach. Figure 5: Comparison of NMTC Investment Structures: [See PDF for image] Source: GAO. [A] Investors in a CDE cannot redeem any of the original QEI during the 7-year period while they are allowed to claim the credit. However, equity investors can receive a return on their investment in a CDE, in the form of dividends or partnership income, for example. [End of figure] The leveraged investment structure may make NMTC investment more attractive to some investors because it allows investors to invest in the CDE who may not be able to claim tax credits but could still benefit from the economic returns. The investment structure can be used to separate the tax benefits of the investment from the economic benefits of the investment. For example, an investment fund partnership makes a $1 million leveraged qualified equity investment in a CDE where $400,000 of the money comes from the equity investors in the partnership and the other $600,000 comes from a bank as an interest- only loan to the investment partnership with a balloon payment after 7 years. The CDE that receives the QEI reinvests the money by loaning "substantially all" of the $1 million to a QALICB. In this structure, the economic and tax benefits are separated: the bank receives interest payments on the loan to the CDE and, after 7 years, the bank will also be entitled to collect principle payments on the loan while the equity investors are entitled to claim the NMTC for 7 years, totaling 39 percent of the total $1 million QEI--not just the $400,000 that was originally invested as equity. NMTC equity investors may also receive a return on their investment, in the form of dividends or partnership income, for example, during the 7-year period while they can claim the credit. However, neither the investment fund partnership nor the underlying investors can redeem any portion of the QEI during this period and still remain eligible to claim the credit. The leveraged investment structure may also offer a more attractive combination of risk and return than direct investment. From the bank's perspective in the example above, this investment structure may be attractive because the loan-to-value ratio[Footnote 22] is more favorable than it would have been if the debt was not being combined with the investors' equity. The more favorable ratio may compensate the bank for assuming a greater degree of risk, most notably if the business that receives the loan from the CDE defaults on its loan agreement. In that case, the bank's investment is only secured by the equity in the original investment partnership ($400,000 in the example above). From the equity investor's perspective, if the business defaults on its loan, they are still allowed to claim the full amount of the credit--as long as the business that receives the funds is a qualifying business in the year the loan is made. As the NMTC program has grown, investors have used more complicated investment structures, such as tiered investments. According to CDFI Fund data, 81 percent of investors making NMTC investments through December 2006 used tiered (including both equity and leveraged) NMTC investment structures, with investors in more recent years being more likely to use tiered structures. For example, 69.1 percent of investors making QEIs in 2003 and 2004 used tiered structures, while 87.5 percent of investors making QEIs in 2005 and 2006 used tiered structures. The CDFI Fund's Data Collection Systems Are Operational: The CDFI Fund uses four data collection systems to administer and monitor the NMTC program. All of these data collection systems were operational before they were needed to collect data and to help the CDFI Fund monitor NMTC compliance. These data collection systems include (1) the Allocation Agreement System (AAS), (2) the Allocation Tracking System (ATS), (3) the Community Investment Impact System (CIIS), and (4) the New Markets Compliance Monitoring System (NCMS). Figure 6 illustrates how the AAS, ATS, and CIIS, combine to populate the NCMS, which the CDFI Fund uses to monitor CDEs' compliance with their allocation agreements. Figure 6: Interaction of CDFI Fund NMTC Data Collection Systems: [See PDF for image] Source: The CDFI Fund and GAO. [End of figure] A brief description of these data collection systems follows. * The AAS contains information on the allocation agreements that CDEs enter into with the CDFI Fund. The AAS was operational as of August 2003 and is primarily used by the CDFI Fund's legal staff to ensure that NMTC contracts are properly executed. * The ATS is the primary system that the CDFI Fund uses to monitor QEIs that have been made and track CDEs (allocatees), suballocatees, and investors in the CDEs. The ATS contains information reported by the CDEs on the type of QEI that is made in the CDE, the amount of the investment, the CDE that received the investment, whether the CDE that initially received the allocation transferred the allocation to a suballocatee, and how much of the allocation was transferred. In addition, the ATS contains data reported by CDEs on the equity investors in the NMTC program. The ATS was operational as of November 2003. * The CIIS collects information about CDEs and the investments that they make in low-income communities. CIIS data is collected through two reports: the Institution Level Report (ILR) and the Transaction Level Report (TLR). The ILR provides information on the CDEs, as well as their loan purchases and Financial Counseling and Other Services (FCOS) activities, and the TLR provides information the CDEs' loans and investments in QALICBs and in other CDEs. The CIIS began receiving data in May 2004. * The NCMS combines data from the CIIS, ATS, AAS, and other CDFI Fund data collection systems and is used to monitor whether CDEs remain compliant with their allocation agreements. CDFI Fund officials said that the NCMS has been operational since April 2005 and that the system was in place in time to allow the CDFI Fund to monitor first round allocatees' compliance with their respective allocation agreements. Financial Institutions and Individuals Are the Primary NMTC Investors, and CDEs Most Often Use NMTC Investments to Make Loans to Qualified Businesses: Banks and individuals constitute the majority of NMTC claimants when qualified equity investments are originally made.[Footnote 23] Taken together, banks and individuals accounted for 70 percent of NMTC claimants through 2006. Banks and other corporations that invested in the credit had relatively large net assets. Individuals who invested in the NMTC had, on average, higher incomes than other taxpayers. The CDEs applied for far more allocation dollars than were available, receiving only about 11 percent of $107 billion in allocation authority for which they applied. The CDEs made investments in low-income communities, most often in the form of term loans to businesses.[Footnote 24] The businesses that received these loans used them for a variety of purposes but chiefly to finance new commercial real estate construction and rehabilitation. The communities where the investment projects were located were dispersed across states, and about 90 percent of projects were located in areas designated as "areas of high distress" because of factors such as low median incomes and high unemployment rates, including businesses in highly distressed areas, such as federally designated Empowerment Zones and Enterprise Communities. NMTC Investors Tend to Be Financial Institutions with Larger Net Assets and Individuals with Higher Incomes: Although the NMTC program has attracted a variety of types of investors, as table 2 indicates, banks and individuals make up the majority of investors, accounting for 70 percent of NMTC investors. Other corporate investors, such as real estate development firms and insurance companies, and still other types of investors, including estates and trusts, make up the remainder of investors in the CDEs. Banks and other regulated financial institutions also account for the majority of NMTC investment funds. Table 2: NMTC Claimant Types: Investor type: Bank or other regulated financial institution; Number of claimants: 155; Percent of claimants: 37.8. Investor type: Individual investor; Number of claimants: 132; Percent of claimants: 32.2. Investor type: Other corporate investor; Number of claimants: 76; Percent of claimants: 18.5. Investor type: Other; Number of claimants: 47; Percent of claimants: 11.5. Investor type: Total; Number of claimants: 410; Percent of claimants: 100.0. Source: GAO analysis of CDFI Fund data. [End of table] Corporations and individuals that claim the tax credit differ from other taxpayers in several key ways. Corporations investing in the NMTC tend, on average, to have larger total assets. For example, the average total assets for corporations that made NMTC investments was $98.3 billion in tax year 2003, while the average total assets[Footnote 25] for all corporations was $9.9 million (the average total assets for banks, the most common type of corporate NMTC claimant, was close to $990 million in 2003).[Footnote 26] Similarly, individual NMTC investors had larger adjusted gross incomes than other individuals who filed tax returns in tax year 2003.[Footnote 27] The average adjusted gross income for individual NMTC investors was about $1.2 million, while the average income for all individual taxpayers was about $47,600. In response to our survey, NMTC investors indicated that they decided to participate in the NMTC program for a variety of reasons.[Footnote 28] As table 3 shows, our investor survey revealed that the majority of NMTC investors indicated that the ability to claim the tax credit (over 75 percent of investors) and obtain a return on their investment (82 percent of investors) played at least a moderately important role in their decision to make an NMTC investment. Investors also indicated that improving conditions in low-income communities (90 percent) and creating and retaining jobs (78 percent) were at least moderately important motivations. About 40 percent of investors also noted that the credit played an important role in helping them remain compliant with other government regulations. Table 3: Reasons NMTC Investors Invested in the NMTC Program (in Percentages): Reason: Obtain the tax credit; Very great to moderate: 76.7 (67.9, 84.1); Little or no extent: 23.3 (15.9, 32.1). Reason: Obtain return on investment; Very great to moderate: 82.1 (73.8, 88.6); Little or no extent: 17.9 (11.4, 26.2). Reason: Improve conditions in low-income communities; Very great to moderate: 90.1 (82.7, 95.1); Little or no extent: 9.9 (4.9, 17.3). Reason: Comply with government regulations; Very great to moderate: 41.2 (33.4, 48.9); Little or no extent: 58.8 (51.1, 66.6). Reason: Create or retain jobs; Very great to moderate: 77.8 (69.1, 85.0); Little or no extent: 22.2 (15.0, 30.9). Reason: Expand lending relationships with special purpose borrowers; Very great to moderate: 52.0 (42.9, 61.0); Little or no extent: 48.0 (39.0, 57.1). Source: GAO survey of NMTC investors. Note: Numbers in parentheses indicate confidence intervals. [End of table] Over time the number of new investors in the CDEs that receive NMTC allocations has increased. For example, 19 percent of investors that made their first QEIs in 2003 were new investors. The CDFI Fund defines new investors as investors making their first investment in a particular CDE. The percentage of new investors increased with investment made in 2004 through 2006 to a high of 69 percent in 2006 (through mid-December). Most investors that have participated in the NMTC program have only made one qualified equity investment. However, CDFI Fund data indicate that it is not uncommon for NMTC investors to participate in more than one QEI. For example, as of December 2006, about 55 percent of NMTC investors have only participated in one QEI, while 33 percent of NMTC investors participated in from two to five QEIs and 12 percent of investors participated in five or more QEIs. The Average Expected Return on NMTC Investment Has Declined: As NMTC investment structures have become increasingly complex in recent years, the expected rate of return for NMTC investments decreased.[Footnote 29] NMTC investments made in 2003 had an average expected rate of return, which includes any return on the equity investment and the tax credit, of 8.2 percent while investments in later years had an average expected rate of return of only 6.8 percent. This decline could be a result of the greater perceived risk for investments made at the beginning of the program. According to CDFI Fund officials, as the program has developed and investors have gained a better understanding of the manner in which the credit can be used, investors' perceived risk in making NMTC investments has likely declined. A factor contributing to the decline may be that as table 4 shows, NMTC investors reported that they have become more familiar with the operations and investment portfolios of the CDEs they invested in after making NMTC investments. Table 4: Investor Knowledge of CDE Operations (in Percentages): Level of knowledge: Before investing in NMTC; Very high to moderate extent: 48.9 (39.6, 58.2). Level of knowledge: After investing in NMTC; Very high to moderate extent: 93.4 (85.5, 97.7). Source: GAO survey of NMTC investors. Note: Numbers in parentheses indicate confidence intervals. [End of table] However, even though the reported expected rate of return on NMTC investments has fallen, investors indicate that they remain concerned about the market risk of NMTC investments and the possibility that businesses that receive NMTC investments could default on their loans. For example, our investor survey indicates that an estimated 86 percent (78.2, 92.0) of investors said that they were at least moderately concerned that their investment would not achieve its expected rate of return, and 81 percent (71.8, 87.9) of investors said that they were at least moderately concerned that the business that received their NMTC investment would default on its loan. CDEs Apply for More NMTC Allocations Than Are Available and Relatively Few CDEs Receive Allocations: For all allocation rounds combined, CDEs have applied for over $107 billion in NMTC allocation and received only about 11 percent of requested allocation dollars. As table 5 shows, the percentage of dollars awarded in relation to the dollars requested has remained fairly constant during the four allocation rounds, but in each round CDEs have applied for far more in NMTC allocations than the CDFI Fund has had the authority to award based on the NMTC's authorizing legislation. The amount awarded as a percentage of the amount requested varied by at most 6 percentage points over the rounds. In general, CDEs applied for more in allocation authority in rounds where larger amounts were available for allocation. Table 5: NMTC Allocations Awarded by Round: Dollars in billions. Round 1 (2003); Amount requested: $26.0; Amount awarded: $2.5; Percentage awarded: 9.6. Round 2 (2004); Amount requested: 30.4; Amount awarded: 3.5; Percentage awarded: 11.5. Round 3 (2005); Amount requested: 22.9; Amount awarded: 2.0; Percentage awarded: 8.7. Round 4 (2006); Amount requested: 28.3; Amount awarded: 4.1; Percentage awarded: 14.5. Total; Amount requested: $107.6; Amount awarded: $12.1; Percentage awarded: 11.2. Source: GAO analysis of CDFI Fund data. [End of table] For all allocation rounds combined, the CDFI Fund received 1,078 NMTC applications from CDEs and only 223, or about 22 percent, received allocations. As table 6 shows, between 19 percent and 25 percent of CDEs that applied for allocations received them in each round. Table 6: CDEs That Applied for NMTC and Received Allocations by Round: Round: Round 1 (2003); Number of applicants: 345; Number receiving allocations: 66; Percentage receiving allocations: 19. Round: Round 2 (2004); Number of applicants: 271; Number receiving allocations: 63; Percentage receiving allocations: 23. Round: Round 3 (2005); Number of applicants: 208; Number receiving allocations: 41; Percentage receiving allocations: 23. Round: Round 4 (2006); Number of applicants: 254; Number receiving allocations: 63; Percentage receiving allocations: 25. Round: Total; Number of applicants: 1,078; Number receiving allocations: 233; Percentage receiving allocations: 22. Source: GAO analysis of CDFI Fund data. [End of table] CDFI Fund officials indicated that NMTC applications will score particularly well to the extent that, among other things, the applicants commit to: (1) providing products with particularly flexible or nontraditional rates and terms; (2) serving severely economically distressed communities, including communities that have been targeted for redevelopment by other governmental programs; and (3) investing more than the minimally required 85 percent of NMTC proceeds into low- income communities. We observed the application reviewer training session in 2005 and noted that the CDFI Fund encouraged application reviewers to pay particular attention to types of projects and financing terms being proposed in the applications. One example we noted was that CDFI Fund officials instructed NMTC application reviewers to base a portion of each application's overall score on the commitment of the applicant to serve highly economically distressed areas. Businesses Primarily Receive Loans from CDEs That They Use Chiefly for Investment in Commercial Real Estate: CDEs that received NMTC allocations have used their allocations to make investments totaling $3.1 billion through fiscal year 2005, primarily in the form of loans to businesses in low-income communities. According to CDFI Fund data, these loans are used chiefly for constructing and rehabilitating commercial real estate and are also used to purchase fixed assets for businesses[Footnote 30] and to provide working capital for businesses.[Footnote 31] For example, these loans have been used to finance a range of activities, such as the rehabilitation of historic buildings and the operation of mixed-use real estate development. Other uses include the construction or operation of cultural arts centers, frozen pizza manufacturing, and the construction of charter schools. As figure 7 shows, about 75 percent of the dollar value of these loans and investments was used for investment in commercial real estate. Figure 7: NMTC Loans and Investment by Type of Activity for Fiscal Years 2003 through 2005: [See PDF for image] Source: GAO analysis of CDFI Fund data. [End of figure] According to data reported by CDEs to the CDFI Fund, most investment (88 percent) made by the CDEs in businesses comes in the form of term loans.[Footnote 32] According to CDFI Fund data, the most common types of loans being made to qualifying business with better rates and terms[Footnote 33] come in the form of loans with below market interest rates (80 percent of reported NMTC dollars) and lower-than-standard loan origination fees (56 percent of reported NMTC dollars). As figure 8 illustrates, other types of favorable financial packages that qualifying business take advantage of include things like interest-only loans, loans with longer-than-standard amortization periods, and higher loan-to-value ratios than are traditionally required. Figure 8: NMTC Dollars Used in Loans with Better Rates and Terms: [See PDF for image] Source: GAO analysis of CDFI Fund data. [End of figure] The Communities Receiving the Investment Tend to Be More Highly Distressed: Through their allocation agreements with the CDFI Fund, all allocatees are required to use at least some portion of their allocation to serve designated "areas of higher distress,"[Footnote 34] which may have a greater need for economic development funds than areas that meet the NMTC program's minimal requirements. For example, 51 percent of projects serve areas with a median income of less than 60 percent of area median income, and 47 percent of projects serve areas with unemployment rates at least 1.5 times the national average. In addition, over one-fourth of NMTC projects are located in federally designated Empowerment Zones and 51 percent of all NMTC projects are in Small Business Administration-designated Historically Underutilized Business Zones. NMTC projects are distributed across states. Activities reported through fiscal year 2005 included 583 projects, located in 45 states, the District of Columbia, and Puerto Rico. Table 7 shows the top 10 states organized by the total dollar amount of NMTC investment and the total number of projects. Appendix III contains the full list of the number of NMTC projects by state. Table 7: Top 10 States by NMTC Dollars through Fiscal Year 2005: Top 10 states by amount of dollars invested: California; Total dollar amount of loans and investment: $303,081,270; Number of NMTC projects: 58. Top 10 states by amount of dollars invested: New York; Total dollar amount of loans and investment: 239,178,566; Number of NMTC projects: 25. Top 10 states by amount of dollars invested: Ohio; Total dollar amount of loans and investment: 201,857,969; Number of NMTC projects: 69. Top 10 states by amount of dollars invested: Maine; Total dollar amount of loans and investment: 153,527,250; Number of NMTC projects: 13. Top 10 states by amount of dollars invested: Wisconsin; Total dollar amount of loans and investment: 149,131,108; Number of NMTC projects: 26. Top 10 states by amount of dollars invested: Missouri; Total dollar amount of loans and investment: 146,165,868; Number of NMTC projects: 22. Top 10 states by amount of dollars invested: Massachusetts; Total dollar amount of loans and investment: 145,059,237; Number of NMTC projects: 34. Top 10 states by amount of dollars invested: Kentucky; Total dollar amount of loans and investment: 135,117,406; Number of NMTC projects: 44. Top 10 states by amount of dollars invested: North Carolina; Total dollar amount of loans and investment: 126,420,590; Number of NMTC projects: 14. Top 10 states by amount of dollars invested: Washington; Total dollar amount of loans and investment: 125,703,680; Number of NMTC projects: 19. Source: GAO analysis of CDFI Fund data. [End of table] NMTC Investors Report That the NMTC Increases Investment in Low-Income Communities and Statistical Analysis Indicates That These Investments May Be Financed by Shifting Assets from Other Uses and Some New Investment: The results of our investor survey and statistical analysis indicate that the NMTC may be increasing investment in eligible low-income communities by participating investors, which is consistent with the program's purpose. Increased investment in low-income communities can occur when NMTC investors increase their total funds available for investment or when they shift funds from other uses. One limitation with our survey is that NMTC investors responding to our survey, because they benefit from claiming the credit, have an interest in ensuring that the NMTC program continues to operate. Our survey indicated that most NMTC investors increased the share of their investment budget for low-income communities because of the credit. However, in many cases the survey also indicated that the credit alone may not have been sufficient to justify the investment and meeting other government regulations may be an important incentive for making NMTC investments. In addition, about two-thirds of investors also indicate that NMTC investors have a track record of investing in low- income communities, which may mean that some investment was shifted from other low-income community investments. Our statistical analysis suggests that corporations investing in the NMTC are shifting investment funds while individuals who make NMTC investments may be increasing their overall level of investment. Neither our statistical analysis nor the results of our survey allow us to determine definitively whether shifted investment funds came from higher income communities or from other low-income community investments. A complete evaluation of the NMTC program's effectiveness requires determining whether the program's economic and social benefits to low- income communities offset its costs, which include costs such as forgone tax revenue and economic distortions evidenced by shifting investment funds. We did not conduct this complete evaluation for this report because sufficient data were not available. The CDFI Fund is currently working with a contractor to develop plans for a comprehensive program evaluation, which may include some aspects of program effectiveness. NMTC Investors Reported That They Increased Their Investment in Low- Income Communities Because of the Credit, but Other Factors May Also Play a Role: In response to our survey, most NMTC investors said that they would probably or definitely not have made the same investment with the same terms if they had not been eligible to claim the credit. An estimated 88 percent of investors said that they would not have made the same investment without the NMTC. Of these investors who would not have made the same investment without the NMTC, 75 percent (66.6, 82.7) also indicated that in the absence of the NMTC they would not have made a similar investment in the same community. Moreover, 64 percent (54.9, 72.5) of investors said that they increased the share of their investment budget that is designated for low-income communities because of the NMTC. Most NMTC investors have experience in low-income community investment. Nearly two-thirds of investors have additional investment in low-income communities that does not qualify for the NMTC. Sixty-one percent (53.2, 69.4) of respondents currently had additional investments in low- income communities that were not eligible QEIs, and 29 percent of investors had made one or more investments in other CDEs or similar organizations that mainly serve low-income communities but cannot be used to claim the NMTC. This interest in low-income community investment is also reflected in survey responses where 90 percent of investors said the goal of improving conditions in low-income communities influenced their decision to invest in the NMTC from a moderate to very great extent. Most investors also indicated that they plan to make additional NMTC investments. The survey responses indicate that in many cases, the credit alone may not have been sufficient to justify the investment. The NMTC can also be packaged with a number of other government incentives to make the investment more attractive. About half of respondents combine the NMTC with at least one other government incentive that can provide additional tax benefits to the investor. As figure 9 shows, state and local tax abatements are the most popular type of government incentive used. Some respondents that packaged the NMTC with other government incentives indicated that their ability to package the credit played an important role in their decision to make the investments, which may indicate that in some cases, the NMTC, in and of itself, is not a strong enough incentive to encourage investment in low-income communities. Figure 9: NMTC Investors Packaging the NMTC with Other Government Incentives: [See PDF for image] Source: GAO. Note: Confidence intervals for each of the categories in the figure are as follows: Historic Rehabilitation Tax Credits (19.6, 35.5), Historic Rehabilitation Easement Deduction (.1, 5.8), Brownfields tax incentive (7.2, 20.1), Empowerment Zone/Enterprise Community funding (15.7, 31.8), and State/local tax abatements (28.5, 46.1). [End of figure] Meeting other government regulations may also be an important incentive for making NMTC investments. Over 40 percent of the investors reported that they use the NMTC to remain compliant with the Community Reinvestment Act (CRA), which rates depository institutions on their record of helping to meet the credit needs of their entire community. Seventy-one percent (58.3, 80.8) of investors that are required to comply with the CRA use their NMTC investment to help meet their CRA obligations. For investors using the NMTC to meet CRA requirements, 94 percent (83.4, 98.8) view it as very or somewhat important in their decision to make the investment. Nearly half of NMTC investors also reported that they make investments eligible for the Low-Income Housing Tax Credit, a tax credit for investment in rental housing targeted to lower income households. However, less than one-half of the investors that also invest in the Low-Income Housing Tax Credit view it as an alternative to the NMTC. One explanation for this is that these investors may be making other low-income community investments as a means for complying with government requirements such as the CRA. For example, of the survey respondents that participated in both the NMTC and the Low-Income Housing Tax Credit, nearly three quarters of these investors are also required to comply with the CRA. Statistical Analysis Suggests That Some NMTC Investment by Participating Investors May Be Investment Shifted from Other Assets and Some May Be New Investment: Our statistical analysis of corporations and individuals that claimed the NMTC indicates that some NMTC investment may be shifted from other uses and some investment could be new investment. Statistical analysis of corporations that claimed the NMTC indicates that, in general, NMTC investment funds are not new investment made from an increase in total funds available. When combined with information from the survey, this statistical result may indicate that corporations are shifting NMTC investment funds from other uses. Statistical analysis of individuals who invested in the NMTC indicates that in the aggregate, NMTC investment funds represent, at least in part, an overall increase in investment levels. Because corporate NMTC investment accounts for the majority of QEIs, the increased investment associated with participation in the program is likely to come primarily from funds shifted from other uses. Statistical analysis of corporations that claimed the NMTC indicates that NMTC investment funds are not likely to represent new overall investment. To assess whether NMTC investments represent new funds, we compared the growth rate in net assets of corporations that made NMTC investments to the growth in net assets of a similar group of corporations that did not make NMTC investments over time. We selected our comparison group using a stratified random sample of taxpayers based on total assets at the end of the tax period. We drew the comparison groups based on 2000 tax year data because this was the year before the credit could be claimed and in that year we would not expect any changes in behavior because of the credit. If NMTC investments represent new investment funds[Footnote 35] then we would expect the net assets of NMTC participants to grow faster over time than the net assets of corporations that did not make NMTC investments. Using multiple specifications, our results suggest that corporate claimants' net assets are not growing faster than similar corporations that did not make NMTC investments.[Footnote 36] Rather than new investment, NMTC investment could represent a shift of investment by participating corporations from high-or moderate-income communities to low-income communities. This conclusion follows from combining evidence from the survey of investors with evidence from the statistical analysis.[Footnote 37] Because our analysis does not show a faster growth rate for NMTC investors, it is possible that the credit has no effect on investor behavior, but instead rewards investors for investment in low-income communities that would have been made in the absence of the credit. However, the effect of the credit may also be to shift investment from other low-income communities or from high-or moderate-income communities. Although it contains some contrary indicators about the effect of the credit, the survey of investors that benefit from claiming the credit, indicated that most investments would not have occurred in the absence of the credit and that NMTC investors had increased their investments in low-income communities because of the credit. Therefore, we infer that the most likely effect of the credit is that it shifts investment by participating investors into low- income communities from higher income communities. Further analysis of the components of net assets, total assets, and total liabilities, which are discussed in appendix II, produced inconclusive results regarding the source of the shifted funds. Shifted investment funds, in contrast to new investment funds, indicate that investors are decreasing investment in another asset or assets by some or all of the amount that they invest in the NMTC program. Investors might choose to shift funds for a variety of reasons, including a higher rate of return expected from the NMTC investment, a need to make an investment eligible for meeting CRA requirements, or the ability to establish new business relationships. Regardless of the reason, if funds are shifted as the result of a tax benefit, the shifting potentially creates other economic costs, including the opportunity cost of other uses of the funds, and benefits.[Footnote 38] These costs and any benefits that accrue to low-income communities should also be considered when evaluating the overall effectiveness of the tax credit in addition to the revenue costs of the program. When analyzing the effect of NMTC participation on the net assets of corporations, our results consistently showed no effect. Further, when we tested our results using different data specifications, we were still not able to detect an effect. However, our analysis of the NMTC's effect on net assets for corporations had several limitations. For example, the amount of NMTC investment might be small enough relative to a corporation's total size that our statistical models could fail to detect a positive effect of the NMTC investment on corporations' asset levels. We attempted to mitigate this problem by basing our analysis on firm-level data, the smallest unit of analysis available, and growth in assets over time. In addition, we did not have data for total liabilities. We calculated a corporation's total liabilities by subtracting stockholders' equity and retained earnings from the "total liabilities and shareholders' equity" line-item on the tax return. Additionally, our data made it difficult to identify which industry NMTC corporate investors participated in, another variable that would have helped strengthen our analysis. Similar analysis of individuals who invested in the NMTC indicates that at least some portion of their investment may represent an overall increase in investment (or "new" investment) rather than investment shifted from other uses. To assess whether NMTC investments represent new funds, we compared the wealth of individuals who made NMTC investments to the wealth of a similar group of individuals who did not make NMTC investments over time. If NMTC investments represent new investment funds[Footnote 39] then we would expect the wealth of NMTC claimants to grow faster over time than the wealth of nonclaimants. As table 8 shows, the NMTC is associated with a positive effect on the growth in NMTC investors' wealth. This means that NMTC investors' wealth is growing at a faster rate than similar investors who did not make NMTC investments.[Footnote 40] Thus, according to our analysis for individual NMTC claimants, the NMTC program investors appear to be increasing their investment in low-income communities because their QEIs represent investments that they would not have made otherwise and these investments are placed into low-income communities according to program rules.[Footnote 41] The increase in wealth for individuals can be broken down into its components, such as interest-bearing assets and business assets. The NMTC can have indirect effects on these components of wealth through its effect on after tax income. In addition to potentially producing ordinary returns on investment (such as dividend payments), part of the return on NMTC investments comes in the form of reduced tax liabilities. Because they are paying less in taxes, NMTC investors have more income available for investing in other types of assets and for consumption. As table 8 shows, our results are consistent with individuals placing at least a portion of this income into interest- bearing assets, such as savings accounts or certificates of deposit. As table 8 shows, these new NMTC assets also appear to take the form of business assets, including partnerships. Increases in business assets may be consistent with typical NMTC investment structures where many individuals are investing through pass-through entities. Table 8: Effects of NMTC Individual Investor Participation on Wealth: Measure: Wealth; Effect on growth of:[A]: Positive[B]. Measure: Interest-bearing assets; Effect on growth of:[A]: Positive. Measure: Dividends; Effect on growth of:[A]: None. Measure: Real estate; Effect on growth of:[A]: None. Measure: Business; Effect on growth of:[A]: Positive. Source: GAO analysis of IRS and CDFI Fund data. [A] Results are based on comparisons of the 2000 to 2004 growth in each category, comparing growth for NMTC investors and an individual tax return representing the closest match among non-NMTC claimants from our comparison group. [B] For categories where a positive effect is identified, our analysis was statistically significant at the 5 percent level. [End of table] In our analysis, NMTC participation by individuals was associated with greater growth in wealth, and most variables measuring this association were highly statistically significant. In addition, various checks that we performed were consistent with the results we present above. However, as was also the case with our analysis of corporate investors, several data limitations exist for our analysis of individual investors. For instance, we did not have direct data on asset holdings. Consequently, we estimated wealth based on income streams reported on tax returns. In addition, some assets are particularly difficult to measure. Business assets are especially susceptible to measurement errors as income streams from these assets may vary widely from year to year. This means that assets not generating reportable returns, such as stock holdings that do not generate dividends in a particular year, do not appear in our estimates for that year. We have attempted to mitigate this problem by conducting a series of tests, such as using a 3-year average of wealth and asset variables, to confirm the consistency of our results. These tests and data limitations are discussed in more detail in appendix II. Further Analysis Is Needed to Determine Whether the Economic Costs of Shifting Investment Are Justified by Any Economic and Social Benefits to Low-Income Communities: A complete evaluation of the program's effectiveness goes beyond identifying whether the credit increases investment in low-income communities by participating investors and also requires determining both whether non-NMTC investors would have made the same investments that the NMTC investors made if the NMTC investors had not made the investment and whether the program's benefits to low-income communities offset its costs, which include costs such as forgone tax revenue and potential economic inefficiencies created by shifting investment funds. Fully examining the effectiveness of the NMTC requires addressing at least two main issues: where do NMTC investment funds come from and do NMTC investments generate economic benefits in low-income communities? Because of data limitations, the relative youth of the NMTC program, and the inherent difficulties of measuring program costs and benefits, a full evaluation is beyond the scope of this report. However, our finding that the NMTC program causes claimants to shift their investment portfolios suggests that the program might generate some additional economic costs, such as the opportunity cost of redirecting investment resources from other, potentially valuable uses.[Footnote 42] Whether these economic costs are justified depends on the economic benefits that are generated in the low-income communities and the extent to which these benefits accrue to the targeted population. This highlights the importance of assessing the benefits of the program in eligible communities so that one can assess whether the costs are justified by the benefits of the program. The CDFI Fund has hired a contractor to design a comprehensive study to evaluate the NMTC program. The study design will be completed by mid- 2007, and the study will begin after the design is complete. During the design phase, the contractor will complete five case studies of NMTC investments. The study could potentially evaluate the effect that the NMTC is having on factors such as job creation and economic growth in areas that receive the credit. These issues fell outside the scope of this report. IRS and the CDFI Fund Monitor NMTC Compliance, but Additional Opportunities Exist to Better Measure Noncompliance and Identify NMTC Investors: IRS monitors CDEs' compliance with NMTC laws and regulations, and IRS is conducting a compliance study but is not yet selecting CDEs to audit in a manner that represents all types of CDEs. The CDFI Fund monitors CDEs' compliance with their allocation agreements through its data collection systems and, on a more limited basis, by making site visits. The CDFI Fund has tested its data systems and developed policies and procedures for site visits. IRS and the CDFI Fund developed a memorandum of understanding (MOU) in an attempt to clarify the roles and responsibilities of both agencies in ensuring NMTC compliance, and IRS has access to CDFI Fund data. However, additional efforts could help IRS receive information in a more useful format. In addition to IRS and the CDFI Fund, investors and CDEs play a role in ensuring that CDEs remain compliant and the credit is not recaptured. IRS's Compliance Study Methodology Could Be Improved to Be More Representative of the Population of CDEs: IRS is responsible for ensuring that CDEs and NMTC investors adhere to NMTC laws and regulations. As part of its effort to monitor CDEs' compliance, IRS is conducting a study to monitor CDEs' compliance with NMTC legislative requirements, focusing on CDEs' compliance with the "substantially all" requirement to invest at least 85 percent of their QEIs within 1 year of receiving the investment. IRS officials said that they chose to focus on CDEs' compliance with the "substantially all" requirement because they believed that this was the area where noncompliance with NMTC provisions was most likely to occur. The current compliance study will provide IRS with some information about audited CDEs compliance with the "substantially all" requirement, including information about whether funds were invested in a timely manner and whether the investments were made to qualifying businesses. However, IRS did not select first round NMTC allocatees to audit in a manner that likely represents the full range of CDEs. IRS envisioned that its compliance study would focus on verifying that CDEs were in compliance with statutory requirements through examining CDEs' tax returns and auditing CDEs. IRS has taken steps to develop and implement the compliance study, such as training auditors to conduct NMTC examinations and developing a training manual that provides examiners with background on the NMTC program, key issues to consider when reviewing whether CDEs meet the "substantially all" requirement, and information to familiarize auditors with the investment structures that NMTC investors use to make investments. IRS is currently auditing 20 of the 66 first round allocatees. IRS officials said that they initially planned to conduct examinations of early round CDEs using a sample of CDE tax returns that would yield a valid 95 percent level of confidence for the study's results. IRS expected that all CDEs that received early round allocations would file income tax returns within a year or two of the award date, and that shortly after all the CDEs' tax returns were filed, IRS would have enough returns to select a valid sample that would yield the desired confidence level. However, IRS changed its selection process because it took more time than expected for CDEs to file tax returns, and the volume of returns filed was not sufficient for IRS to draw a valid sample in a timely manner. IRS officials said that the delay for most CDEs occurred because of the lapse of time between the date that the CDE executed agreements with the CDFI Fund and when the CDE actually collected equity investments and began operations. As a result of the delay in acquiring tax returns for its study, IRS modified its overall compliance strategy in two ways. First, it decided to verify that each allocatee filed a tax return as a way to monitor CDEs' filing compliance. IRS intends to continue to monitor CDEs' filing compliance until they are confident that the entities will file as required.[Footnote 43] Second, IRS discontinued the sample approach and decided to manually review every return that it could identify. IRS initially requested over 80 tax returns from tax years 2003 and 2004. Of the returns that IRS had received by June 2006, it chose to facilitate audits of CDEs that filed 2004 tax year returns that had some indication of NMTC activity. According to IRS, because of the delays in when CDEs were awarded NMTC allocations and the time in which they began filing tax returns, IRS did not develop specific criteria for deciding which CDEs to audit. An IRS official said that IRS wanted to start its compliance study as soon as possible and the filing time lags created delays. IRS indicated that IRS will continue with this selection process until it reaches a point where there are sufficient returns placed in the examination stream to produce meaningful results.[Footnote 44] IRS plans to use the results of the compliance study, which will take several years to complete, to guide its future enforcement efforts. While IRS's current compliance study will provide the agency with information about CDEs' compliance with NMTC laws and regulations, the compliance study will have limited value if the audit selection process does not represent the full range of transactions. We have previously reported that taxpayer compliance studies should be representative of the population for which compliance is being measured and reasonably designed for developing compliance measures for the taxpayer population as a whole and for subgroups of taxpayers (such as suballocatees in the case of the NMTC program).[Footnote 45] IRS's current plan for its compliance study could be improved to adhere to these standards more closely. Given IRS's intent to rely on the study to guide enforcement efforts, the results of not having a study representative of the population could be lost tax revenue and increased cost through inefficient use of resources. IRS could change its strategy to make its results more useful as its compliance work progresses. IRS plans to audit 15 to 25 CDEs from each allocation round until it feels that compliance levels warrant a reduced number of audits. While it may be too resource intensive to conduct a statistically valid study with fully generalizable results, IRS could work with the CDFI Fund to develop criteria for determining which CDEs to audit. For example, IRS could use CDFI Fund data to categorize CDEs that invest in different types of projects or CDEs that use different types of investment structures for NMTC purposes. As the program expands and more tax return data are available for future rounds, IRS could use the audit results from its initial CDE audits, along with developing these criteria for identifying which CDEs it will audit, in order to produce compliance study results that will be more representative of the entire population of NMTC allocatees. The CDFI Fund Has Systems and Procedures in Place to Monitor CDEs' Compliance with Allocation Agreements: The CDFI Fund is monitoring CDEs to ensure that they remain compliant with their allocation agreements through the New Markets Compliance Monitoring System (NCMS) and, on a more limited basis, site visits. The CDFI Fund took steps to ensure that its data collection and reporting systems are reliable and valid, such as testing its data collection systems and the interaction between these systems multiple times before using them to identify CDE noncompliance. These steps help to reasonably ensure that the CDFI Fund data are adequately maintained and properly disclosed in reports. CDFI Fund databases rely on data that CDEs self-report to the CDFI Fund. However, the CDFI Fund has several mechanisms in place, such as providing written instructions to CDEs on how to report data and providing a help desk for CDEs to call when they have questions about reporting information to the CDFI Fund, that help ensure that the data they collect are accurate and reliable. In addition, data used to populate the NCMS are subject to several validity checks to ensure accuracy. CDFI Fund officials have also conducted a limited number of site visits to CDEs, one goal of those site visits being to ensure that data are being accurately reported. Our review of the CDFI Fund's NCMS system and site visits indicates that the CDFI Fund has instituted policies and procedures that should allow it to collect the information that it believes it needs to meet its compliance program's objectives of identifying CDEs that are no longer compliant with their allocation agreements. According to our Government Auditing Standards, agencies should develop internal controls, including controls that will ensure that programs operate effectively and efficiently and that data collected are reliable and valid.[Footnote 46] The CDFI Fund uses the NCMS to detect allocatees' noncompliance with their allocation agreements relating to authorized uses of NMTC allocations, restrictions on the use of NMTC allocations, and other special provisions that are included in an allocation agreement. If the NCMS identifies a CDE as being out of compliance with its allocation agreement, the CDFI Fund contacts the allocatee to let it know that the NCMS has identified it as noncompliant. The CDFI Fund officials then attempt to determine why the CDE is noncompliant and take steps necessary to bring the CDE back into compliance with the terms of its allocation agreement. As of January 2007, the CDFI Fund had identified nine CDEs that were not compliant with their allocation agreements and one CDE that was not in compliance with the NMTC program's "substantially all" requirement. For example, in one case the CDFI Fund determined through data reported in the NCMS that the CDE was serving communities that were outside its approved service area. In this case, the areas that the CDE was investing in still qualified for NMTC investment. In response, the CDFI Fund amended the CDE's allocation agreement by expanding the CDE's service area. Six of the noncompliance CDEs were first round allocatees that had not, as required in their allocation agreements, issued 60 percent of their QEIs by the end of September 2006. The CDFI Fund is working with most of these allocatees to correct the problem; however, one first round allocatee has had its NMTC allocation revoked and another CDE returned its allocation as a result of not meeting this requirement. In the case where the CDFI Fund used the NCMS to identify a CDE that was failing the "substantially all" test, the CDFI Fund referred the problem to the IRS. In this case, the CDE was able to correct the problem within 6 months, the amount of time CDEs are given to correct failing the "substantially all" test, and further action was not required. The CDFI Fund developed policies and procedures for conducting site visits to CDEs where CDFI Fund officials check the validity of data reported by CDEs' to the CDFI Fund and obtain additional information about CDEs' efforts to remain compliant. These policies and procedures include criteria for prioritizing which allocatees warrant a site visit, the key information items to collect on a site visit, and a plan for using the information after the site visit is complete. As of November 2006, the CDFI Fund had conducted four site visits, two in 2005 and two in 2006, and indicated that it intends to conduct more visits in the future. A CDFI Fund official indicated that the CDFI Fund has plans to conduct three site visits in fiscal year 2007. So far, the CDFI Fund has visited one multiyear allocatee, one CDE that the NCMS had identified as noncompliant, a CDE that participates in other CDFI Fund programs, and a bank that received an allocation award. The process of conducting a site visit goes through several steps. A site visit can be triggered when a CDE meets one or more of the seven criteria established by the CDFI Fund, which include whether the NCMS identified the CDE as noncompliant and whether the allocatee received awards in multiple allocation rounds. Once the CDFI Fund contacts the allocatee it intends to visit, CDFI Fund officials review the data that the CDE reported to the CDFI Fund and identify any areas of concern that the CDFI Fund will investigate during the site visit. During the visit, CDFI Fund officials review other documents, such as board meeting minutes and financial documents, and conduct interviews with key staff members. CDFI Fund officials also review documentation that the CDE maintains in order to ensure that the data the CDE reported to the CDFI Fund are accurate and reliable. After the site visit is complete, CDFI Fund officials prepare a site visit report using information gathered before and during the site visit. If the CDFI Fund does not find the CDE to be in default with its allocation agreement, no further enforcement action is taken. However, if the initial CDFI Fund report finds that the CDE is not compliant with its allocation agreement, the report is passed on to CDFI Fund senior management who then either approve or disapprove the report's finding. While these site visits do not yield generalizable results, they do supplement the information that the CDFI Fund receives through the NCMS. Unlike IRS, which must audit CDEs to determine if they are compliant with the NMTC's laws and regulations, the CDFI Fund is able to use data reported by CDEs as its primary mechanism for reviewing CDEs' compliance with their allocation agreements. As a result, the CDFI Fund is able to use data in the NCMS in conjunction with site visits that do not yield generalizable results in order to detect when a CDE is no longer compliant with its allocation agreement. If a CDE is determined to be noncompliant, the CDFI Fund can restrict the CDE's access to the NMTC program. According to CDFI Fund officials, if they find a "serious occurrence of noncompliance," such as a CDE failing to perform any of the transactions that it agreed to perform, the CDE would be found in default. To the extent possible, the CDFI Fund would assist the CDE in correcting the areas in which it was determined to be noncompliant--this could include amending or modifying the CDE's allocation agreement. If the CDE is not able to come back into compliance, the CDFI Fund could potentially bar that CDE from future allocation rounds, or if the CDE has not yet issued all its QEIs, the CDFI Fund could revoke its ability to make additional investments using its current allocation. Thus far, the CDFI Fund has not had to take these actions against any CDE as a result of the outcome of site visits. IRS and the CDFI Fund Have an MOU for Compliance Monitoring, but Additional Opportunities Exist to Monitor Investor Compliance: IRS and the CDFI Fund have cooperated in their compliance efforts. As part of their response to our initial NMTC report,[Footnote 47] the CDFI Fund and IRS developed an MOU in an effort to clarify the roles and responsibilities of both with respect to monitoring NMTC compliance. IRS and the CDFI Fund have had additional discussions to identify ways for the CDFI Fund to streamline the data that it provides to IRS. While IRS and the CDFI Fund have worked together to monitor NMTC compliance, the two agencies could collect additional information that would help the IRS monitor compliance by NMTC investors, an area where neither the CDFI Fund nor IRS has chosen to dedicate resources. According to the MOU completed in 2004, the CDFI Fund is responsible for carrying out the NMTC program's application and allocation procedures. In addition, the MOU states that the CDFI Fund will permit designated IRS staff to have access to CDFI Fund databases, provide IRS with the relevant findings and assessments of any site visits to NMTC allocatees conducted by CDFI Fund staff, and notify IRS of any potential credit recaptures. Also, on behalf of IRS, the CDFI Fund also includes compliance questions that CDEs respond to in its database regarding recapture and investments that CDEs have made in low-income communities. If the CDFI Fund determines from the answers to these questions that the CDE may be in danger of having the NMTC recaptured, it is to forward the information to IRS. According to the terms of the MOU, IRS is responsible for the collection and determination of any tax as deemed appropriate. In addition, the MOU notes that IRS is responsible for establishing processes and procedures to ensure that taxpayers are in compliance with the NMTC's tax provisions, and IRS will provide the CDFI Fund with quarterly information, to the extent permitted by law, regarding any CDEs that fail to meet the NMTC's legal requirements. IRS and the CDFI Fund have identified data sharing as an area where their cooperation could be improved. While IRS has access to CDFI Fund data, according to IRS officials, they have had difficulty selectively obtaining the information that they are most interested in from the CDFI Fund's data systems. According to IRS officials, a more streamlined format for sharing data between IRS and the CDFI Fund would allow IRS to better target noncompliance. CDFI Fund officials said that they are working with IRS to develop a streamlined compliance data report, and they indicated that IRS has been cooperative in working with them. An IRS official agreed that the two agencies are working together to develop a more user-friendly data report specifically for IRS. IRS is also taking steps to increase the amount of information available about NMTC investors. IRS is in the process of finalizing a new form that will require CDEs to report to IRS the amount of QEI that NMTC investors made at the investment's original issue. IRS currently does not have these data for all claimants because the CDFI Fund data that IRS currently uses to identify credit claimants does not track claimants in cases when the underlying QEI is sold to another investor. In addition, IRS is finalizing a second form that will require CDEs to notify the original equity investor in an NMTC investment if the credit is being recaptured. With these forms and the CDFI Fund data, IRS will have a complete record of the initial NMTC investors in a CDE and how much they invested. However, further steps could be taken to identify NMTC investors and ensure that only eligible taxpayers claim the credit and that they claim the correct amounts. NMTC investors are allowed to sell their equity share in a CDE, which determines their NMTC eligibility, to other investors after the initial investment has taken place, and neither the IRS nor the CDFI Fund tracks NMTC investors after the original investment. IRS officials indicated that the forms they are finalizing cannot be used to track the selling of an investor's equity share in a CDE because they will not be refiled if the investment is sold to another investor after the original investment. As a result, IRS and the CDFI Fund will not be able to identify all NMTC investors and the amount of QEI that they made if an investor's equity share in a CDE is sold after the original investment. When evaluating other tax credits, we have noted that IRS is responsible for ensuring that taxpayers claim those tax credits for which they are entitled.[Footnote 48] If IRS and the CDFI Fund developed ways to identify investors and the amounts they invested, even when NMTC investors sell their equity shares in a CDE, they would be better able to ensure that credits are claimed correctly. Our analysis of IRS and CDFI Fund data indicates that many NMTC investments may be sold after the original QEI is made in the CDE, making it difficult for IRS to identify all eligible NMTC claimants and the amounts that they are eligible to claim. When we compared potential tax credit claimants in IRS's databases to claimants in the CDFI Fund's database, we noted that more investors were identified as being eligible to claim the credit in IRS's taxpayer data than in the CDFI Fund's data on claimants when a QEI is originally issued. According to IRS, requiring individual investors to report sales of NMTC investments could place an undue burden on taxpayers. However, IRS told us that this would be useful information for its compliance monitoring efforts--both for identifying investors eligible to claim the NMTC on their tax returns and for identifying tax credit investors if IRS is forced to recapture the credits from investors when a CDE is no longer compliant with the "substantially all" requirement.[Footnote 49] The CDFI Fund already collects information from CDEs in its database identifying the initial investors and how much NMTC eligible investment has been made by investors that did not participate in tiered equity or leveraged NMTC transactions. Further, a NMTC investor with prior experience investing in CDEs and a representative of a CDE said that in their experience, CDEs are already able to identify subsequent holders of NMTC qualified equity investments when one NMTC investor sells its equity share in a CDE to another investor, and CDEs could potentially be able to report that information to the CDFI Fund or IRS. In the case where investors in a partnership that has NMTC investments sell their share in the partnership, it may be more difficult for CDEs to identify who the correct tax credit claimants would be, although the CDE would still know which partnerships own QEI in the CDE. Currently, neither IRS nor CDFI Fund data make it possible to identify completely who is eligible to claim the tax credit and how much they are entitled to claim. As more NMTC investments are being resold and complicated investment structures are becoming more common, limits on IRS's ability to monitor investor compliance could make IRS vulnerable to a loss of tax revenues caused by taxpayer noncompliance, fraud, and abuse, and it could become increasingly difficult for IRS to identify tax credit claimants if it is forced to recapture the credit. If CDEs reported more complete information about initial NMTC investors and subsequent sales of the equity shares in the CDE that are linked to NMTC eligibility to the IRS or the CDFI Fund, IRS would have better information to track investor compliance. Investors in CDEs Play a Role in Ensuring NMTC Compliance: Investors that responded to GAO's NMTC survey indicated that they are concerned about the possibility of the credit being recaptured and that they play an active role in ensuring that CDEs remain compliant with the laws and regulations that apply to the NMTC program. An estimated 82 percent (74.0, 89.0) of our survey respondents indicated that they are "moderately" to "very highly" concerned about the possibility that the credit could be recaptured. Nearly all investors, 97 percent, reported that they make some effort to ensure that CDEs remain compliant so that the investors avoid recapture. About 72 percent of the survey respondents said that they have regular discussion with CDEs, and 84 percent said they receive regular reports from CDEs. Nearly one-quarter of NMTC investors said that they audit the CDEs in which they made NMTC investments. Figure 10 shows the activities that NMTC investor survey respondents undertake to monitor CDE compliance. Figure 10: Activities Investor Survey Respondents Undertake to Monitor CDE Compliance: [See PDF for image] Source: GAO. Note: Confidence intervals for the data presented in this graphic are as follows: receive regular reports from the CDE (75.8, 90.3), receive independent audit reports about the CDE (74.4, 89.9), have regular discussions with the CDE (62.8, 80.6), make regular site visits to the CDE (33.8, 52.5), created the CDE and staff it ourselves (25.3, 42.3), audit the CDE (16.1, 32.5). [End of figure] Conclusions: The purpose of the NMTC program is to encourage investment and development in low-income communities. Our analysis indicates that the program may be accomplishing part of that objective. In our investor survey, most participating investors said that they increased investment in low-income communities because of the credit. The statistical analysis also showed an increase in investment, with individuals adding new investment and corporations shifting funds from other uses. However, some of the survey evidence may be less consistent with the credit increasing investment (e.g., the prior experience of most NMTC investors with low-income community investment) and, because of data limitations, our statistical evidence may only establish an association between the credit and increased investment, not that the program causes the increase. In any case, the indication that the program increases investment is not sufficient to support conclusions about the program's effectiveness, nor is the fact that the credit shifts investment an indicator of a lack of effectiveness. For example, more information is needed about the economic and social benefits that the low-income communities receive from the investment. This information is only now likely to be available given that the program's implementation was delayed. IRS and the CDFI Fund are implementing a compliance monitoring system in the context of a program that is growing and that is attracting investors that use increasingly complex and sophisticated investment structures. As IRS moves forward with its NMTC compliance study, more rigorous development of criteria for selecting which CDEs to audit could help it better identify the most common compliance issues facing CDEs. Additionally, more complete information on who is eligible to claim the tax credit and the amounts that they are eligible to claim would be useful to IRS in helping ensure that only eligible taxpayers claim the NMTC, and a complete list of eligible NMTC claimants would assist IRS should the IRS need to recapture NMTCs. Recommendations for Executive Action: To ensure IRS is reviewing the full range of NMTC transactions and that the conclusions of its compliance study are more representative of all CDEs with NMTC allocations, we recommend that IRS use CDFI Fund data and the results of its current NMTC compliance study to develop criteria for selecting which CDEs to audit as part of its future compliance monitoring efforts. Additionally, to ensure that eligible taxpayers claim the correct amount of NMTC on their tax returns and IRS is able to identify all tax credit claimants in the event of the credit being recaptured, we recommend that IRS work with the CDFI Fund to further explore options for cost effectively monitoring investor compliance and developing a way to identify NMTC claimants, even in instances where the original investor sells its equity share in a CDE, and the amount of QEI that each investor made. Agency Comments: We received written comments on a draft of this report from the Acting Director of the CDFI Fund and the Commissioner of Internal Revenue; their comments are reprinted in appendices IV and V. Both the IRS and the CDFI Fund agreed with our recommendations. We also incorporated technical corrections to the draft report that we received from both IRS and the CDFI Fund where appropriate. In its response to the draft report, the CDFI Fund characterized GAO's study as indicating that the NMTC has been a highly successful tool for increasing the flow of investments into low-income communities. While our findings do suggest that the NMTC appears to increase investment by participating investors in low-income communities, we also note that further information is needed to fully assess the effectiveness of the NMTC program. We are sending copies of this report to the interested congressional committees, the Commissioner of Internal Revenue, the Director of the Community Development Financial Institutions Fund, and other interested parties. We will make copies available to others on request. In addition, the report will be available at no charge on the GAO web site at [Hyperlink, http://www.gao.gov]. If you or your staff have any questions on matters discussed in this report or would like additional information, please contact me at (202) 512-9110 or at brostekm@gao.gov. Major contributors to this report are acknowledged in appendix VI. Signed by: Michael Brostek: Director, Tax Issues: List of Committees: The Honorable Max Baucus: Chairman: The Honorable Charles E. Grassley: Ranking Minority Member: Committee on Finance: United States Senate: The Honorable John F. Kerry: Chairman: The Honorable Olympia J. Snowe: Ranking Minority Member: Committee on Small Business and Entrepreneurship: United States Senate: The Honorable Christopher Dodd: Chairman: The Honorable Richard S. Shelby: Ranking Minority Member: Committee on Banking, Housing, and Urban Affairs: United States Senate: The Honorable Charles B. Rangel: Chairman: The Honorable Jim McCrery: Ranking Minority Member: Committee on Ways and Means: House of Representatives: The Honorable Nydia M. Velázquez: Chair: The Honorable Steve Chabot: Ranking Minority Member: Committee on Small Business: House of Representatives: The Honorable Barney Frank: Chairman: The Honorable Spencer Bachus: Ranking Minority Member: Committee on Financial Services: House of Representatives: [End of section] Appendix I: Objectives, Scope, and Methodology: Based on consultations with staff at cognizant congressional committees, the objectives of this report are to (1) describe the status of the New Markets Tax Credit (NMTC) program; (2) profile the characteristics of NMTC investors, the Community Development Entities (CDE) that receive NMTC allocations, and the businesses and communities that receive NMTC investments; (3) assess how effective the NMTC has been in bringing new investment to low-income communities by the investors that have participated in the program; and (4) assess the steps that the Internal Revenue Service (IRS) and Community Development Financial Institutions (CDFI) Fund are taking to ensure CDEs and investors are complying with the NMTC and evaluate how effective these steps have been. In order to accomplish these objectives, we used a number of methods of analysis. We met with officials from the CDFI Fund and IRS. We collected documents on the program status and efforts to monitor NMTC compliance. We also analyzed data from the CDFI Fund on the CDEs and their investment in low-income communities and tax return data from tax years 1997 through 2004 for investors in the NMTC program. We used these data to report summary statistics that profile the participants in the program and to conduct statistical analysis that measures the effect of the NMTC on investment. We also surveyed investors in the NMTC program in order to provide additional information on the effect of the credit and characteristics of the investors. To evaluate investment in the CDEs by NMTC investors, we used data from the CDFI Fund's Allocation Tracking System (ATS) on investments reported through mid-December 2006. We used the ATS data to report on the type and size of qualified equity investment (QEI) made in the CDE and the CDE that received the investment. We also used the ATS to analyze the equity investors in the NMTC program. To report on qualified low-income community investments (QLICI) from the CDE to the corresponding qualified active low-income community business (QALICB) we analyzed data from the Community Investment Impact System (CIIS). Specifically, we used data from the CIIS Transaction Level Report (TLR) for fiscal years 2003 through 2005, which provides information on each transaction made as part of a QLICI. To assess the reliability of the ATS and the TLR data sources, we reviewed the CDFI Fund's data quality control procedures and subsequently determined that the data were sufficiently reliable for our purposes. We also reviewed tax data on NMTC investors from IRS's Individual Returns Transaction File (IRTF) and Business Returns Transaction File (BRTF). We identified NMTC claimants using data on original claimants (at the time the QEI was made) from the CDFI Fund's ATS and used their tax return information to determine how NMTC investors differ in size from all taxpayers. In cases where we could not locate a corporation's tax return because the NMTC investor was a subsidiary of a larger parent corporation, we used IRS's National Account Profile to link the subsidiary to its parent corporation. In these cases, the parent corporation's tax return was used in our analysis. In addition, because original claimants may sell their investment, and along with it their NMTC credit, we identified further claimants as those individuals or corporations that indicated they were eligible to claim the NMTC on their tax returns. This information came from IRS's IRTF or BRTF on the New Markets Tax Credit Form (Form 8874) or as part of the General Business Credit (Form 3800). To assess the reliability of the IRS data sources, we reviewed the IRS's data quality control procedures and subsequently determined that the data were sufficiently reliable for our purposes. To obtain information from investors on the effectiveness of the NMTC, we designed and implemented a Web-based survey to gather information on the investors' motivations and methods. We used CDFI Fund data and interviews with investors to determine the proper points of contact for NMTC investors. Our survey population consists of NMTC claimants and their proxies for cases in which the individual claimant was not principally responsible for deciding to make the NMTC investment. In some cases, one person was designated as the contact point for a group of investors responding to the survey. The survey asked a combination of questions that allowed for open-ended and close-ended responses. Because some investors invested with more than one CDE and because not all investors participated in tiered or leveraged investment structures, the instrument was designed with skip patterns directing investors to comment only on the prepopulated CDE and type of investment structure that they utilized. Therefore, the number of survey respondents for each question varied depending on the number of CDEs in which the investor made a QEI and whether the investor had used tiered or leveraged structures. We pretested the content and format of the questionnaire with knowledgeable investors. During the pretest, we asked the investors questions to determine whether (1) the survey questions were clear, (2) the terms used were precise, (3) the questionnaire placed an undue burden on the respondents, and (4) the questions were unbiased. We also assessed the usability of the Web-based format. We received input on the survey from a CDFI Fund official and made changes to the content and format of the final questionnaire based on pretest results. The survey was conducted using self-administered electronic questionnaires posted on the World Wide Web. We sent e-mail notifications to investors beginning on August 2, 2006.[Footnote 50] We then sent each potential respondent a unique password and user name by e-mail to ensure that only members of the target population could participate in the appropriate survey. To encourage respondents to complete the questionnaire, we sent e-mail messages to prompt each nonrespondent approximately 2 weeks and 3 weeks after the initial e- mail message. We also arranged for contract callers to do phone follow- ups from September 6 to September 8, 2006. We closed the survey on October 3, 2006. Because we attempted to collect data from every investor in the population, there was no sampling error. However, the practical difficulties of conducting any survey may introduce errors, commonly referred to as nonsampling errors. For example, differences in how a particular question is interpreted, the sources of information available to respondents, how the responses were processed and analyzed, or the types of people who do not respond can influence the accuracy of the survey results. We took steps in the development of the surveys, the data collection, and the data analysis to minimize these nonsampling errors and help ensure the accuracy of the answers that were obtained. A second, independent analyst checked all the computer programs that processed the data. The response rate for this survey was 51 percent. We conducted a nonresponse bias analysis by looking at the response rates for eight cells defined by the four types of investors surveyed (financial institutions, individuals, nonfinancial corporations, and other) and the size of the investor's total assets (in the case of corporations) or adjusted gross income (for individuals). We collected this information primarily from the investor's most recent tax return filed with IRS. In cases where we could not identify a tax return (primarily because the corporation had recently been acquired or merged with another corporation) we relied on public information on the corporation's total assets from its most recent annual report. Investors were placed in one of two size categories, either less than the median or greater than the median. Individuals with adjusted gross income less than the median for individuals using the NMTC had the highest response rate at 63 percent followed by financial institutions with a response rate of 56 percent for financial institutions with income above the median and 53 percent for financial institutions with income below the median. Individuals with incomes above the median had the lowest response rate at 32 percent. Differential response rates across analytic subgroups raise the possibility of nonresponse bias. If the respondents provided different responses than the nonrespondents, the survey estimates would be biased. We have weighted the respondents by type and income to reduce this source of nonresponse bias. Unfortunately, there may be other sources of nonresponse bias that we are unaware of and unable to adjust for. A statistician used the data on size and type of investor to create weights that allowed us to project the survey responses to the entire population by assuming that the nonrespondents would have answered the questions as the respondents did. We have treated the respondents as a stratified, random sample and calculated sampling errors as an estimate of the uncertainty around the survey estimates. Ninety-five percent confidence intervals are given in parentheses after the estimates. We are 95 percent confident that each of the confidence intervals in this report will include the true values in the study population. We also used IRS tax data to develop statistical analysis that measures the effect of the NMTC on investment and addresses the question of whether NMTC investments represent new or shifted funds. Using the tax returns of NMTC investors as determined from CDFI Fund and IRS data (see above) we used a multistage sampling methodology to draw a comparison group of tax returns. These methods are more fully described in appendix II. To develop our statistical methodology, we relied on academic journal articles and interviewed experts in the research fields of individual savings and wealth and corporate taxation. To study the effectiveness of the steps that IRS and the CDFI Fund are taking to ensure CDEs and investors are complying with the NMTC and the effectiveness of these measures, we met with officials from the CDFI Fund and IRS. We also collected documents on the program status and efforts to monitor NMTC compliance. We performed our work at GAO Headquarters and the IRS office in New Carrollton, Maryland, from July 2006 through December 2006 in accordance with generally accepted government auditing standards. [End of section] Appendix II: Description of Data and Methodology for Statistical Analysis of the Effect of NMTC Participation on Investment: This appendix describes our data and methodology for assessing whether participation in the NMTC program affects investment by NMTC investors in low-income communities. The NMTC program may affect investment by increasing the overall level of investment (i.e., creating "new" investment) or by causing NMTC investors to shift investment from other uses to investment eligible for the credit. The methodology that we use to detect these changes in investment follows the methodology used in the retirement savings literature. This literature generally compares the wealth or financial assets of participants in retirement savings plans to that of nonparticipants to detect any effect of participation on savings.[Footnote 51] In our assessment of the NMTC program, we compare the wealth or assets of NMTC program participants to that of a group of similar nonparticipants to detect any effect on investment. Our statistical analysis of the effectiveness of the NMTC program in stimulating investment depends on the distinction between new and shifted investment. If our analysis detects new investment, this outcome is consistent with program goals because it may indicate increased investment in low-income communities that would not have occurred in the absence of the credit. If we do not detect new investment, it is possible that the credit has created no change in behavior and investors are just receiving a subsidy for investments that they would have made anyway, which is not consistent with the goals of the program. However, the investment could also be shifted from other communities. The implications for the effectiveness of the program in the case of shifted investment are more ambiguous. It could mean that (1) the credit has induced investors to shift investments from assets invested in other low-income communities, which means that although the credit has generated investments in projects that would not have occurred otherwise, it has not increased investment in low- income communities, or (2) NMTC investments represent funds shifted from higher income communities. The first outcome is not consistent with the NMTC's broader goal of increasing investment in low-income communities as a whole. The second outcome is more consistent with program goals because, as with new investment, it may indicate increased investment funds available to low-income communities. Finally, in the case of both new and shifted investment, NMTC investment may reduce investment by non-NMTC investors (called crowding out) which is also inconsistent with the broader goal of the program. Our data and methodology do not allow us to detect crowding out, and for this reason, we confine our analysis to the effect of the credit on the investment behavior of participants in the NMTC program. A limitation of our statistical analysis is that in the case of no detected change in the overall level of investment, we cannot distinguish between the possible types of shifting or between shifting and the possibility that there has been no change in investment behavior. However, if we combine evidence from our survey of investors with evidence from our statistical analysis, our analysis may provide some indication that the effect of the program on investment in low- income communities by NMTC investors is shifted investment. The survey of investors that benefit from the tax credit indicated that most investments would not have occurred in the absence of the credit (inconsistent with the notion that the credit has no effect on investor behavior), and that NMTC investors had increased their investments in low-income communities because of the credit (inconsistent with the first shifting outcome above). Therefore, we use the second shifting outcome described above to interpret our statistical results in cases where we detect no overall increase in the level of investment by NMTC investors. Description of Data: We identified NMTC investors using both CDFI Fund data and IRS data. We collected data on original claimants (at the time the QEI was made) from the CDFI Fund. We also identified investors from IRS's Returns Transaction File data as those claiming a positive amount for the credit on their tax returns in tax years 2001 through 2004. There were differences in the number of claimants identified from the two different sources with the IRS data resulting in more investors. The source of these differences is unclear as they could indicate incomplete CDFI Fund data, missing taxpayer identification numbers (TIN) in the CDFI Fund data, or a large turnover in credits. In the latter case, investors may not be responding to the incentives of the credit themselves but to the terms constructed by the original investor. However, this is not necessarily the case as some investors we spoke with that had purchased the credit from the original investor indicated that they intended to participate but that the original investor was necessary due to timing issues. Because of the uncertainty over which set of investors is the most relevant for our analysis, we estimated results using both the full sample (IRS and CDFI Fund claimants) and CDFI Fund claimants only. Our conclusions were the same for both groups; however, we are only reporting results for the full sample of NMTC investors identified in IRS and CDFI Fund databases. Our analysis of these data indicated that NMTC claimants were generally higher income (individuals) or had higher total assets (corporations) than the average taxpayer. This prompted us to identify our basic comparison group using a stratified random sample of taxpayers based on adjusted gross income for individuals and total assets at the end of the tax period for corporations. We oversampled high income and total asset taxpayers relative to an unstratified random sample from the same populations. We used quintiles to stratify our sample and drew a random sample of about 4,000 returns per quintile.[Footnote 52] We chose our quintiles and drew the comparison groups based on 2000 tax year data because this was the year before the credit could be claimed and in that year we would not expect any changes in behavior due to the credit. For individuals, we collected all available data from Form 1040 and information from Schedules C and F to form a panel of taxpayers for tax years 1997 through 2004. The data include more than 24,000 individual tax filers and about 80 percent of filers (including NMTC investors) are in the panel for all 8 years. For corporations, we used income data from Form 1120 and balance sheet data from Schedule L to form a panel of corporate taxpayers for tax years 1997 through 2004. These data include more than 14,000 corporate tax filers and about 56 percent of corporate filers were present in at least 7 years. (Forty-eight percent were present for all 8 years and 57 percent of NMTC investors were in all years.) Both individual and corporate NMTC investors were identified using TINs contained in CDFI Fund data and the New Markets Tax Credit Form (Form 8874) or as part of the General Business Credit (Form 3800) in the IRS data. The total number of NMTC claimants identified from these sources was 753. We also estimated asset values for individuals because, unlike IRS balance sheet data for corporations, the IRS data for individuals were limited to income streams and did not include asset levels. We followed the methods used in the Survey of Consumer Finances (SCF) to estimate asset holdings using income streams and rates of return.[Footnote 53] We also expanded on the SCF approach by using more sophisticated modeling to develop estimates of home equity. Rather than attribute to each household the median home value within its income group (as the SCF does), we estimated home equity using the November 1999 Wave (12) of the 1996 Survey of Income and Program Participation. Our controls included total income, age, marital status, and region.[Footnote 54] We then applied these estimated coefficients to tax return information on total income, age, filing status, and region of residence to generate estimates of home equity for each household using 2000 tax data. Negative values were set to zero, and the consumer price index (CPI) (research series)[Footnote 55] was used to adjust the year 2000 estimates for earlier and later years.[Footnote 56] Effects of NMTC Program Participation on Investment: We assessed the effects of NMTC participation by comparing the level of assets and growth in assets of NMTC participants with the level and growth in assets of corporations and individuals that did not participate in the NMTC program.[Footnote 57] We used regression techniques to compare the level of assets of NMTC investors and the relevant comparison group. The results of these models indicate whether the assets of NMTC claimants are higher than those of our comparison group controlling for other individual and corporate characteristics. However, it is possible that this approach is simply picking up the likelihood that NMTC claimants systematically have higher assets than their counterparts (despite our efforts to choose an appropriate comparison group using a stratified random sample). Therefore, we used several methods, including regression and propensity score techniques, to compare the growth of assets over time.[Footnote 58] Differences in growth rates between NMTC investors and the comparison group do not depend on differences in the level of assets. Our baseline model for corporate investors is a fixed effects model of the following form[Footnote 59]: Y(it) = X(it)Beta + µ(it): For corporate investors, Y(it) represents the log of total assets, total liabilities, or net assets; X(it) represents control variables which include the lag of net assets, the NMTC participation dummy, year dummies, and region dummies; and µ(it) represents a random error term.[Footnote 60] Additional control variables are not used because they are included in the fixed effect. These variables include corporate-level characteristics, such as industry, that do not change over time. Statistical analysis of this baseline model indicates that corporate NMTC investment funds are more likely to represent investment funds shifted from other uses. Although there was some evidence that NMTC investors have higher levels of net assets than those in our comparison group, this result was not robust over different specifications of the model. On the other hand, our analysis of growth rates showed no statistically significant effect of NMTC investment status on the growth of net assets. This result means that NMTC investors are not investing at rates different from non-NMTC investors. Unlike the case of asset levels, this result was robust across several specifications involving regression and propensity score methods, as indicated in table 9.[Footnote 61] In addition, the result was qualitatively the same for each quintile, when we used only years 2001 through 2004 in the analysis, when we used median regression, and when our analysis included only banks. Table 9: Growth in Net Assets Using Fixed Effects Regression and Comparisons Based on Nearest Neighbor Propensity Score Matching: Specification[A]: Nearest neighbor matching; Average effect: - 2,883.972; Coefficient: [Empty]; Standard error: 6,740.992. Specification[A]: Matching - CDFI Fund indentified investors; Average effect: -19,000.000; Coefficient: [Empty]; Standard error: 41,265.206. Specification[A]: Baseline regression - full sample; Average effect: ; Coefficient: 0.009; Standard error: 0.091. Specification[A]: Quintile one; Average effect: ; Coefficient: 0.002; Standard error: 0.012. Specification[A]: Quintile two; Average effect: ; Coefficient: 0.020; Standard error: 0.090. Specification[A]: Quintile three; Average effect: ; Coefficient: - 0.065; Standard error: 0.614. Specification[A]: Quintile four; Average effect: ; Coefficient: - 0.065; Standard error: 0.306. Specification[A]: Quintile five; Average effect: ; Coefficient: 0.662; Standard error: 1.134. Specification[A]: Banks only; Average effect: ; Coefficient: 0.082; Standard error: 0.061. Specification[A]: Tax years 2001 through 2004; Average effect: ; Coefficient: -0.100; Standard error: 0.133. Specification[A]: Median regression; Average effect: ; Coefficient: 0.000; Standard error: 0.003. Source: GAO analysis of IRS and CDFI Fund data. Note: No coefficients were significant at the 10 percent or better level. [A] Unless otherwise noted the dependent variable is the difference in the inverse hyperbolic sine of net assets. This transformation was used instead of the related log transformation to better address zero and negative values. [End of table] Further analysis included using instrumental variables for predicting participation in the NMTC. However, we did not find important differences between participants and nonparticipants based on location and participation in other general business credits. We concluded that the problem of endogeneity[Footnote 62] may not be a significant issue for corporations because corporate participants are likely to be exposed to a similar set of investment options as nonparticipants and individual corporate characteristics that affect participation are captured in the fixed effect. We also attempted to identify the source of the shifted investment funds by dividing net assets into components, total assets and total liabilities. However, these results were inconclusive as they were not consistent enough to reach any strong conclusions. A limitation of our analysis of corporations is that the amount of NMTC investment might be small relative to a corporation's total size. This means that our statistical models could fail to detect a positive effect of the NMTC investment on corporations' asset levels even if such an effect exists. We attempted to mitigate this problem by analyzing firm-level data, the smallest unit of analysis available, and growth in assets over time. We assessed the effect of NMTC participation on level and growth of assets for individuals in a manner similar to the analysis for corporations. Our baseline model for individual investors is a fixed effects model of the following form: y(it) = N(it) + X(1it)Beta + V(it): where y is the dependent variable, wealth, for household i at time t; N is an indicator for NMTC investment (which is endogenous, i.e., correlated with the error term); X is a set of exogenous control variables; and are coefficients; and it is an error term. However[Footnote 63], unlike the analysis of corporate investors, we analyzed the effect of NMTC on individuals by estimating an instrumental variable[Footnote 64]s version of the baseline model to account for possible endogeneity of the NMTC participation variable. We concluded that this problem is likely to be worse for individuals than for corporations because individuals are less likely to have the same information about the various business tax incentives so that the decision to participate is not random and likely to be correlated with other explanatory variables. We chose as our instrumental variables the dollar amount of allocation in the state of residence and the presence of other general business credit[Footnote 65]s. These variables are likely to be highly correlated with NMTC participation but not with levels of household wealth. To implement the instrumental variables model, we first estimated N as follows: Nit = X1it + X2it + it: where X2 contains our instrumental variables and the other variables are defined as in the baseline model. This regression is used to predict NMTC participation using presence of a general business credit deduction and the cumulative NMTC allocation in state of residence as instrumental variables.[Footnote 66] We then estimated the baseline fixed effects model with Yit as the log of wealth and Xit as control variables, which include balance due, an NMTC participation dummy (instrumented), year dummies, and region dummies.[Footnote 67] In order to test the effect of NMTC participation on the components of wealth, we also ran regressions with Yit as the log of business assets, real estate assets, dividend assets, and interest bearing assets. Like wealth, these asset levels were measured in thousands of dollars and adjusted into constant dollars using the CPI research series. The results of this analysis for asset levels of individuals are presented in table 10. The coefficient for NMTC investor in the wealth column indicates that the log of wealth (in thousands of dollars) is significantly higher than for noninvestors. Table 10: Baseline Analysis: Instrumental Variables Fixed Effects Regressions on the Full Sample: Variable: NMTC investor; Wealth: Coefficient: 2.346; Wealth: Standard error: 0.331; Interest assets: Coefficient: 2.719; Interest assets: Standard error: 0.665; Business assets: Coefficient: 17.287; Business assets: Standard error: 0.950. Variable: Total income (log); Wealth: Coefficient: 0.245; Wealth: Standard error: 0.006; Interest assets: Coefficient: 0.587; Interest assets: Standard error: 0.011; Business assets: Coefficient: 0.360; Business assets: Standard error: 0.016. Variable: Balance due; Wealth: Coefficient: 0.038; Wealth: Standard error: 0.002; Interest assets: Coefficient: -0.031; Interest assets: Standard error: 0.005; Business assets: Coefficient: 0.070; Business assets: Standard error: 0.006. Variable: New England; Wealth: Coefficient: -0.071; Wealth: Standard error: 0.084; Interest assets: Coefficient: 0.269; Interest assets: Standard error: 0.169; Business assets: Coefficient: 0.324; Business assets: Standard error: 0.242. Variable: East North Central; Wealth: Coefficient: -0.242; Wealth: Standard error: 0.078; Interest assets: Coefficient: 0.028; Interest assets: Standard error: 0.156; Business assets: Coefficient: 0.517; Business assets: Standard error: 0.223. Variable: West North Central; Wealth: Coefficient: 0.003; Wealth: Standard error: 0.097; Interest assets: Coefficient: -0.315; Interest assets: Standard error: 0.196; Business assets: Coefficient: 0.363; Business assets: Standard error: 0.279. Variable: South Atlantic; Wealth: Coefficient: -0.134; Wealth: Standard error: 0.063; Interest assets: Coefficient: -0.170; Interest assets: Standard error: 0.127; Business assets: Coefficient: 0.155; Business assets: Standard error: 0.181. Variable: East South Central; Wealth: Coefficient: -0.232; Wealth: Standard error: 0.101; Interest assets: Coefficient: -0.319; Interest assets: Standard error: 0.203; Business assets: Coefficient: 0.437; Business assets: Standard error: 0.289. Variable: West South Central; Wealth: Coefficient: 0.135; Wealth: Standard error: 0.086; Interest assets: Coefficient: 0.301; Interest assets: Standard error: 0.172; Business assets: Coefficient: 0.630; Business assets: Standard error: 0.246. Variable: Mountain; Wealth: Coefficient: -0.125; Wealth: Standard error: 0.081; Interest assets: Coefficient: -0.062; Interest assets: Standard error: 0.162; Business assets: Coefficient: 0.236; Business assets: Standard error: 0.232. Variable: Pacific; Wealth: Coefficient: -0.227; Wealth: Standard error: 0.075; Interest assets: Coefficient: -0.008; Interest assets: Standard error: 0.150; Business assets: Coefficient: 0.103; Business assets: Standard error: 0.214. Variable: Year 1998; Wealth: Coefficient: 0.097; Wealth: Standard error: 0.016; Interest assets: Coefficient: 0.061; Interest assets: Standard error: 0.032; Business assets: Coefficient: 0.068; Business assets: Standard error: 0.046. Variable: Year 1999; Wealth: Coefficient: 0.158; Wealth: Standard error: 0.016; Interest assets: Coefficient: -0.025; Interest assets: Standard error: 0.032; Business assets: Coefficient: 0.099; Business assets: Standard error: 0.046. Variable: Year 2000; Wealth: Coefficient: 0.372; Wealth: Standard error: 0.016; Interest assets: Coefficient: 0.338; Interest assets: Standard error: 0.032; Business assets: Coefficient: 0.259; Business assets: Standard error: 0.046. Variable: Year 2001; Wealth: Coefficient: 0.211; Wealth: Standard error: 0.016; Interest assets: Coefficient: 0.210; Interest assets: Standard error: 0.032; Business assets: Coefficient: 0.270; Business assets: Standard error: 0.046. Variable: Year 2002; Wealth: Coefficient: 0.061; Wealth: Standard error: 0.016; Interest assets: Coefficient: -0.267; Interest assets: Standard error: 0.033; Business assets: Coefficient: 0.405; Business assets: Standard error: 0.046. Variable: Year 2003; Wealth: Coefficient: 0.069; Wealth: Standard error: 0.017; Interest assets: Coefficient: -0.687; Interest assets: Standard error: 0.033; Business assets: Coefficient: 0.569; Business assets: Standard error: 0.047. Variable: Year 2004; Wealth: Coefficient: 0.072; Wealth: Standard error: 0.017; Interest assets: Coefficient: -0.920; Interest assets: Standard error: 0.035; Business assets: Coefficient: 0.506; Business assets: Standard error: 0.050. Variable: Constant; Wealth: Coefficient: 3.383; Wealth: Standard error: 0.055; Interest assets: Coefficient: -3.022; Interest assets: Standard error: 0.110; Business assets: Coefficient: -8.314; Business assets: Standard error: 0.156. Variable: Number of observations; Wealth: Coefficient: 186,241; Wealth: Standard error: ; Interest assets: Coefficient: 186,241; Interest assets: Standard error: ; Business assets: Coefficient: 186,241; Business assets: Standard error: . Variable: Number of groups; Wealth: Coefficient: 24,933; Wealth: Standard error: ; Interest assets: Coefficient: 24,933; Interest assets: Standard error: ; Business assets: Coefficient: 24,933; Business assets: Standard error: . Variable: Overall R-squared; Wealth: Coefficient: 0.157; Wealth: Standard error: ; Interest assets: Coefficient: 0.190; Interest assets: Standard error: ; Business assets: Coefficient: 0.038; Business assets: Standard error: . Source: GAO analysis of CDFI Fund and IRS data. Notes: The dependent variables are in log form and all bold type indicates significance at the 5 percent level. Further, "NMTC investor" is instrumented using total income (log), balance due, allocations in state of residence (log), presence of another general business credit (log), region, and year. The Middle Atlantic is the omitted region and 1997 is the omitted year. [End of table] The coefficients of these regressions should not be used to generate numeric estimates of the magnitude of the effect that the NMTC has on asset levels. In some cases, the fit of our models is poor and it is difficult to estimate the value of some types of assets, in particular business assets. Our results for both the baseline analysis and propensity scoring are intended to illustrate the direction of the effect that the NMTC has on participating individuals' investments. Nonetheless, these results show that NMTC participants have higher levels of wealth and business assets than those in the comparison group after controlling for individual fixed effects, year, region, and tax balance due--a proxy for risk attitudes. These results are consistent across four of five quintiles, using data for years 2001 through 2004 only, and using 3-year averages for the dependent variable. However, it may be that these differences in asset levels are simply picking up the likelihood that NMTC claimants systematically have higher assets that their counterparts. (The summary statistics show that individual NMTC investors have higher asset levels on average than the comparison group despite our use of a stratified random sample where comparison households were chosen based on levels of adjusted gross income in tax year 2000.) Therefore, as an alternative measure of the effect of NMTC participation, we compare the growth in assets between the two groups using closest neighbor propensity score matching to further narrow the comparison group and estimate the effect of NMTC participation on asset growth.[Footnote 68] We used year 2000 data to estimate propensity scores for future participation in the NMTC. The specification for our propensity scoring is as follows: Prob(N=1) = G(XiBeta ): Where N represents any NMTC participation from 2001 through 2004; X includes age, balance due, total income, presence of another general business credit, wage earnings, and dividend earnings; and G( · ) is the cumulative standard normal distribution. We then estimated the effect of NMTC participation on the change in the log of wealth asset levels from 2000 through 2004. Our results show that individuals who participate in the NMTC have higher growth in interest bearing assets, business assets, and wealth which is consistent with the results we obtained for our instrumental variables regressions. For example, the first column in table 11 indicates that the growth in wealth for NMTC investors was significantly higher than that of noninvestors. Table 11: Growth in Assets: Comparisons Based on Nearest Neighbor Propensity Score Matching: Average NMTC effect: 0.881; Wealth: Standard error: 0.152; Interest- bearing assets: Average NMTC effect: 1.554; Interest-bearing assets: Standard error: 0.351; Business assets: Average NMTC effect: 3.112; Business assets: Standard error: 0.527; Real Estate assets: Average NMTC effect: 0.385; Real Estate assets: Standard error: 0.441; Dividend assets: Average NMTC effect: 0.650; Dividend assets: Standard error: 0.431. Source: GAO analysis of IRS and CDFI Fund data. Notes: Average effects are on the difference in the log of the asset from 2000 to 2004 and bold type denotes significance at the 1 percent level. [End of table] Literature Review and Credits: To develop our methodology, we relied heavily on savings literature, which generally compares the wealth or financial assets of participants in retirement savings plans to those of nonparticipants to detect any effect of participation on savings. The following list of publications provided us with important information in developing our methodological approach. 1. Engen, Eric M., and William G. Gale. "The Effects of 401(k) Plans on Household Wealth: Differences Across Earnings Groups." NBER Working Paper No. 8032, 2000. 2. Engen, Eric M., William G. Gale, and John Karl Scholz. "Do Saving Incentives Work?" Brookings Papers on Economic Activity, vol., no.1 (1994). 3. Engen, Eric M., William G. Gale, and John Karl Scholz. "The Illusory Effects of Saving Incentives on Saving." Journal of Economic Perspectives, vol. 10, no. 4 (1996). 4. Hubbard, R. Glenn, and Jonathan S. Skinner. "Assessing the Effectiveness of Saving Incentives." Journal of Economic Perspectives, vol. 10, no. 4 (1996). 5. Pence, Karen M. "401(k)s and Household Saving: New Evidence from the Survey of Consumer Finances." Finance and Economics Discussion Series 2002-6. Washington, D.C.: Board of Governors of the Federal Reserve System, 2002. 6. Poterba, James M., Steven F. Venti, and David A. Wise. "Do 401(k) Contributions Crowd Out Other Personal Saving?" Journal of Public Economics, vol. 58 (1995). 7. Poterba, James M., Steven F. Venti, and David A. Wise. "How Retirement Saving Programs Increase Saving." Journal of Economic Perspectives, vol. 10, no. 4 (1996). 8. Poterba, James M., Steven F. Venti, and David A. Wise. "Personal Retirement Saving Programs and Asset Accumulation: Reconciling the Evidence." NBER Working Paper No. 5599, 1996. We also consulted several experts in the course of our work, including Arthur Kennickel, Karen Pence, James Poterba, and Paul Smith, to discuss the methodology for our statistical analysis. They provided comments that we incorporated into our statistical models. [End of section] Appendix III: NMTC Investment Data by State, Fiscal Years 2003 through 2005: State: California; Total dollar amount of loans and investment: $303,081,270; Percentage of all loans and investment: 9.74; Number of NMTC projects: 58; Percentage of NMTC projects: 9.95. State: New York; Total dollar amount of loans and investment: 239,178,566; Percentage of all loans and investment: 7.68; Number of NMTC projects: 25; Percentage of NMTC projects: 4.29. State: Ohio; Total dollar amount of loans and investment: 201,857,969; Percentage of all loans and investment: 6.49; Number of NMTC projects: 69; Percentage of NMTC projects: 11.84. State: Maine; Total dollar amount of loans and investment: 153,527,250; Percentage of all loans and investment: 4.93; Number of NMTC projects: 13; Percentage of NMTC projects: 2.23. State: Wisconsin; Total dollar amount of loans and investment: 149,131,108; Percentage of all loans and investment: 4.79; Number of NMTC projects: 26; Percentage of NMTC projects: 4.46. State: Missouri; Total dollar amount of loans and investment: 146,165,868; Percentage of all loans and investment: 4.70; Number of NMTC projects: 22; Percentage of NMTC projects: 3.77. State: Massachusetts; Total dollar amount of loans and investment: 145,059,237; Percentage of all loans and investment: 4.66; Number of NMTC projects: 34; Percentage of NMTC projects: 5.83. State: Kentucky; Total dollar amount of loans and investment: 135,117,406; Percentage of all loans and investment: 4.34; Number of NMTC projects: 44; Percentage of NMTC projects: 7.55. State: North Carolina; Total dollar amount of loans and investment: 126,420,590; Percentage of all loans and investment: 4.06; Number of NMTC projects: 14; Percentage of NMTC projects: 2.40. State: Washington; Total dollar amount of loans and investment: 125,703,680; Percentage of all loans and investment: 4.04; Number of NMTC projects: 19; Percentage of NMTC projects: 3.26. State: Minnesota; Total dollar amount of loans and investment: 122,587,357; Percentage of all loans and investment: 3.94; Number of NMTC projects: 13; Percentage of NMTC projects: 2.23. State: Oklahoma; Total dollar amount of loans and investment: 112,092,186; Percentage of all loans and investment: 3.60; Number of NMTC projects: 24; Percentage of NMTC projects: 4.12. State: Oregon; Total dollar amount of loans and investment: 111,464,317; Percentage of all loans and investment: 3.58; Number of NMTC projects: 14; Percentage of NMTC projects: 2.40. State: Maryland; Total dollar amount of loans and investment: 106,171,382; Percentage of all loans and investment: 3.41; Number of NMTC projects: 14; Percentage of NMTC projects: 2.40. State: New Jersey; Total dollar amount of loans and investment: 83,439,000; Percentage of all loans and investment: 2.68; Number of NMTC projects: 7; Percentage of NMTC projects: 1.20. State: Pennsylvania; Total dollar amount of loans and investment: 77,111,177; Percentage of all loans and investment: 2.48; Number of NMTC projects: 21; Percentage of NMTC projects: 3.60. State: Arizona; Total dollar amount of loans and investment: 68,476,055; Percentage of all loans and investment: 2.20; Number of NMTC projects: 8; Percentage of NMTC projects: 1.37. State: District of Columbia; Total dollar amount of loans and investment: 67,715,807; Percentage of all loans and investment: 2.18; Number of NMTC projects: 10; Percentage of NMTC projects: 1.72. State: Texas; Total dollar amount of loans and investment: 65,644,265; Percentage of all loans and investment: 2.11; Number of NMTC projects: 11; Percentage of NMTC projects: 1.89. State: Michigan; Total dollar amount of loans and investment: 57,541,869; Percentage of all loans and investment: 1.85; Number of NMTC projects: 10; Percentage of NMTC projects: 1.72. State: Virginia; Total dollar amount of loans and investment: 55,898,873; Percentage of all loans and investment: 1.80; Number of NMTC projects: 8; Percentage of NMTC projects: 1.37. State: Rhode Island; Total dollar amount of loans and investment: 55,235,675; Percentage of all loans and investment: 1.77; Number of NMTC projects: 3; Percentage of NMTC projects: 0.51. State: Utah; Total dollar amount of loans and investment: 53,884,716; Percentage of all loans and investment: 1.73; Number of NMTC projects: 14; Percentage of NMTC projects: 2.40. State: Georgia; Total dollar amount of loans and investment: 38,516,906; Percentage of all loans and investment: 1.24; Number of NMTC projects: 4; Percentage of NMTC projects: 0.69. State: Florida; Total dollar amount of loans and investment: 38,261,093; Percentage of all loans and investment: 1.23; Number of NMTC projects: 8; Percentage of NMTC projects: 1.37. State: Louisiana; Total dollar amount of loans and investment: 36,162,671; Percentage of all loans and investment: 1.16; Number of NMTC projects: 4; Percentage of NMTC projects: 0.69. State: Connecticut; Total dollar amount of loans and investment: 34,819,477; Percentage of all loans and investment: 1.12; Number of NMTC projects: 3; Percentage of NMTC projects: 0.51. State: Indiana; Total dollar amount of loans and investment: 26,098,460; Percentage of all loans and investment: 0.84; Number of NMTC projects: 3; Percentage of NMTC projects: 0.51. State: Tennessee; Total dollar amount of loans and investment: 22,249,867; Percentage of all loans and investment: 0.71; Number of NMTC projects: 21; Percentage of NMTC projects: 3.60. State: Iowa; Total dollar amount of loans and investment: 20,229,952; Percentage of all loans and investment: 0.65; Number of NMTC projects: 5; Percentage of NMTC projects: 0.86. State: Nebraska; Total dollar amount of loans and investment: 18,778,563; Percentage of all loans and investment: 0.60; Number of NMTC projects: 2; Percentage of NMTC projects: 0.34. State: Delaware; Total dollar amount of loans and investment: 17,000,000; Percentage of all loans and investment: 0.55; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: Mississippi; Total dollar amount of loans and investment: 16,310,758; Percentage of all loans and investment: 0.52; Number of NMTC projects: 2; Percentage of NMTC projects: 0.34. State: Colorado; Total dollar amount of loans and investment: 15,942,664; Percentage of all loans and investment: 0.51; Number of NMTC projects: 7; Percentage of NMTC projects: 1.20. State: Idaho; Total dollar amount of loans and investment: 12,890,000; Percentage of all loans and investment: 0.41; Number of NMTC projects: 10; Percentage of NMTC projects: 1.72. State: Illinois; Total dollar amount of loans and investment: 12,503,895; Percentage of all loans and investment: 0.40; Number of NMTC projects: 8; Percentage of NMTC projects: 1.37. State: Arkansas; Total dollar amount of loans and investment: 10,616,786; Percentage of all loans and investment: 0.34; Number of NMTC projects: 4; Percentage of NMTC projects: 0.69. State: West Virginia; Total dollar amount of loans and investment: 7,398,340; Percentage of all loans and investment: 0.24; Number of NMTC projects: 8; Percentage of NMTC projects: 1.37. State: New Mexico; Total dollar amount of loans and investment: 6,050,000; Percentage of all loans and investment: 0.19; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: Alabama; Total dollar amount of loans and investment: 5,000,000; Percentage of all loans and investment: 0.16; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: South Carolina; Total dollar amount of loans and investment: 3,607,755; Percentage of all loans and investment: 0.12; Number of NMTC projects: 2; Percentage of NMTC projects: 0.34. State: Alaska; Total dollar amount of loans and investment: 3,138,132; Percentage of all loans and investment: 0.10; Number of NMTC projects: 2; Percentage of NMTC projects: 0.34. State: Puerto Rico; Total dollar amount of loans and investment: 1,474,956; Percentage of all loans and investment: 0.05; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: Wyoming; Total dollar amount of loans and investment: 1,461,532; Percentage of all loans and investment: 0.05; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: Nevada; Total dollar amount of loans and investment: 588,750; Percentage of all loans and investment: 0.02; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: Montana; Total dollar amount of loans and investment: 457,200; Percentage of all loans and investment: 0.01; Number of NMTC projects: 1; Percentage of NMTC projects: 0.17. State: Hawaii; Total dollar amount of loans and investment: 250,000; Percentage of all loans and investment: 0.01; Number of NMTC projects: 2; Percentage of NMTC projects: 0.34. Totals; Total dollar amount of loans and investment: $3,112,313,380; Percentage of all loans and investment: 100.00; Number of NMTC projects: 583; Percentage of NMTC projects: 100.00. Source: GAO analysis of CDFI Fund data. [End of table] [End of section] Appendix IV: Comments from the Community Development Financial Institutions Fund: Department Of The Treasury: Community Development Financial Institutions Fund: 601 Thirteenth Street, Nw, Suite 200 South: Washington, DC 20005: Director: January 17, 2007: Mr. Michael Brostek: Director, Tax Issues: U.S. Government Accountability Office: 441 G Street, N.W. Washington, D.C. 20548: Dear Mr. Brostek: Thank you for providing the Community Development Financial Institutions (CDFI) Fund with the opportunity to comment on the draft GAO report, "New Markets Tax Credit Appears to Increase Investment by Investors in Low-Income Communities, but Opportunities Exist to Better Monitor Compliance (GAO-PUB No. 07-296)." The New Markets Tax Credit (NMTC) Program has grown considerably since GAO's initial program report in January of 2004. The CDFI Fund has now made 233 NMTC allocation awards totaling $12.1 billion in allocation authority, which to date has generated over $5.7 billion of investments in Community Development Entities (CDEs) serving low-income communities throughout the country. The tax credit has been used to support a wide variety of community and economic development initiatives including, among others, the financing of charter schools, health care facilities, manufacturing businesses, grocery-anchored retail centers, and numerous other commercial and mixed-use real estate projects. Through FY 2005, the NMTC Program has generated financing for the construction or rehabilitation of over 43 million square feet of real estate, and has helped to create or retain 72,000 construction jobs and 20,000 full time equivalent jobs in businesses in low-income communities. The CDFI Fund appreciates the detailed analysis and valuable recommendations that you have offered in this second program report. We were particularly encouraged by several conclusions you reached as part of your analysis. To cite just a few: * Sixty nine percent (69%) of the investors making investments in CDEs in 2006 had not previously made investments in those entities, suggesting the NMTC is fostering new community development finance relationships. * The average expected returns on NMTC investments have declined from 8.2% to 6.8%, suggesting increased program efficiencies and investor confidence in the NTMC Program. * Communities receiving NMTC investments tend to be more highly distressed than minimally required under program rules. * Investors indicated that most investments would not have occurred in the absence of the credit, and that they had increased their investments in low-income communities because of the credit. * The most likely effect of the credit is that it shifts investment by participating investors into low-income communities from higher-income communities. Taken as a whole, these findings and others in your report suggest that the NMTC has been a highly successful tool for increasing the flow of investments into the nation's most distressed communities. The CDFI Fund concurs with the two Recommendations for Executive Action that you have provided in your report, both of which are related to areas of compliance that fall under the purview of the Internal Revenue Service (IRS). The CDFI Fund will continue to collaborate with IRS to help meet these compliance obligations in the most efficient and cost- effective manner possible. Thank you again for the opportunity to review and comment upon your draft report. We appreciate your efforts and the collaborative relationship that you fostered during the course of your review. Sincerely, Signed by: Peter Dugas: Acting Director: [End of section] Appendix V: Comments from the Internal Revenue Service: Department Of The Treasury: Internal Revenue Service: Washington, D.C. 20224: Commissioner: January 18, 2007: Mr. Michael Brostek: Director, Tax Issues: U.S. Government Accountability Office: 441 G Street, N.W. Washington, D.C. 20548: Dear Mr. Brostek: I have reviewed your draft report titled, "New Markets Tax Credit Appears to Increase Investment by Investors in Low-Income Communities, but Opportunities Exist to Better Monitor Compliance" (GAO-PUB No. 07- 296). I agree with the recommendations in the report, and I have enclosed a detailed response to address your recommendations. As part of our efforts to monitor compliance and ensure adherence to the New Markets Tax Credits (NMTC) laws and regulation, we are conducting a study to monitor compliance with the NMTC legislative requirements. Our focus has been primarily on the Community Development Entities' (CDEs) compliance with the "substantially all" requirement. This area holds the greatest potential for noncompliance with the Internal Revenue Code. Additionally, we continue to redirect our compliance program efforts as information becomes available from the Community Development Financial Institutions (CDFI) Fund, as additional examination results are collected and as new developments or trends are identified in the NMTC environment. In future periods, as more NMTC awards are made and more CDEs become active, our audit efforts will increase accordingly. In the future, increased sharing of the CDFI Fund data and our collection of information from audits will help develop criteria to streamline our examinations. Data sharing will increase our efficiency and ultimately limit the Government's exposure to potential inappropriate tax loss. We will also explore with the CDFI Fund cost effective options to monitor NMTC investor non-compliance. I appreciate the time the GAO Team spent reviewing our initial measures to monitor compliance with the NMTC and the good working relationship during the course of the review. If you have any questions or need additional information on our compliance study, please contact Kelly Cables, Director, Performance Management, Quality Assurance, and Audit Assistance, at (202) 283-8334. Sincerely, Signed by: Mark W. Everson: Enclosure: Recommendation 1: To ensure the IRS is reviewing the full range of NMTC transactions and that the conclusions of their compliance study are more representative of all CDEs with NMTC allocations, we recommend that IRS use CDFI Fund data and the results of its current NMTC compliance study to develop criteria for selecting which CDEs to audit as part of its future compliance monitoring efforts. Response: We agree with your recommendation. The use of CDFI Fund data is a valuable part of our compliance plan for selecting CDEs to audit. Currently, the IRS and CDFI Fund are working together to focus and enhance the informational reports that we currently receive from the CDFI Fund databases as part of our audit selection process. Additionally, as we collect more information from the audits that we have in process, we will be able to identify additional criteria that should be used to select CDEs for audits. This selection process will assure that our tax compliance efforts do not create the opportunity for tax loss nor increase those costs associated with our compliance efforts. In the future, we plan to continue this collaboration with the CDFI Fund to assure that we are able to develop a compliance plan that is cost efficient by using CDFI Fund data. Recommendation 2: Additionally, to ensure that eligible taxpayers claim the correct amount of NMTC on their tax returns and IRS is able to identify all tax credit claimants in the event of the credit being recaptured, we recommend that IRS and the CDFI Fund further explore options for cost effectively monitoring investor compliance and developing a way to identify NMTC claimants, even in instances where the Qualified Equity Investment (QEI) giving rise to the credit is sold, and the amount of QEI that each investor made. Response: We agree with your recommendation. We plan to work with the CDFI Fund to explore options for monitoring investor compliance and identifying NMTC claimants, even in instances where the QEI giving rise to the credit is sold, and for determining the amounts of QEI that each investor made. This would facilitate the identification of NMTC claimants in the event of a recapture and would help to ensure that eligible taxpayers claim the correct amount of NMTC on their tax returns. After these options are identified, we will evaluate them to determine which ones would cost effectively allow compliance monitoring of NMTC claimants and the amounts of QEI they made. Technical Comments: We also provided technical comments from Chief Counsel to GAO on January 5, 2007. [End of section] Appendix VI: GAO Contact and Staff Acknowledgments: GAO Contact: Michael Brostek, (202) 512-9110 or brostekm@gao.gov: Acknowledgments: In addition to the contact named above, Kevin Daly, Assistant Director; Thomas Gilbert; Evan Gilman; Tami Gurley-Calvez; Katherine Harper; Stuart Kaufman; Summer Lingard; Don Marples; Donna Miller; Ed Nannenhorn; Karen O'Conor; and Cheryl Peterson made key contributions to this report. FOOTNOTES [1] Pub. L. No. 106-554 (2000). [2] The original legislation that authorized the program allowed for $15 billion of equity investment to qualify for the NMTC program. However, the Gulf Opportunity Zone Act of 2005, Pub L. No. 109-135 (Dec. 21, 2005) authorized an additional $1 billion of NMTC equity for qualified investments in areas affected by Hurricane Katrina, and Pub. L. No. 109-432 (Dec. 20, 2006) extended the NMTC for an additional year (through 2008) with an additional $3.5 billion of NMTC allocation authority. [3] GAO, New Markets Tax Credit Program: Progress Made in Implementation, but Further Actions Needed to Monitor Compliance, GAO- 04-326 (Washington, D.C.: Jan. 30, 2004). [4] A low-income community is defined as a census tract (1) in which the poverty rate is at least 20 percent or (2) outside a metropolitan area in which the median family income does not exceed 80 percent of median statewide family income or within a metropolitan area in which the median family income does not exceed 80 percent of the greater statewide or the metropolitan area median family income. After October 22, 2004, the Secretary of the Treasury was authorized to issue regulations designating targeted populations that may be treated as low- income communities and procedures for determining which entities are qualified active low-income community businesses with respect to such populations. In addition, the definition of a low-income community included certain areas not within census tracts, tracts with low population, and census tracts with high migration rural counties. [5] GAO-04-326. [6] Community development financial institutions and specialized small business investment companies automatically qualify as CDEs and only need to register as CDEs rather than apply for certification. [7] For more information on how NMTC awards are determined and the criteria that the CDFI Fund uses to select which CDEs will receive allocations, refer to GAO-04-326, pp. 5-9. [8] Beginning in the year the investment is made, investors are entitled to claim the credit for a 7-year period with 5 percent of the investment claimed in each of the first 3 years and 6 percent in each of the last 4 years. Investors are allowed to carry the credit back 1 year and carry the credits forward for a 20-year period. [9] "Substantially all" means that CDEs must use (within 12 months) at least 85 percent of investor proceeds in years 1 through 6 and 75 percent in year 7 of the investment. CDEs can satisfy this requirement by two methods: (1) direct tracing of investments to specific qualified low-income community investments or (2) showing that at least 85 percent of their aggregate gross assets are invested in qualified low- income community investments. [10] For example, an investor may have an interest in beginning a particular NMTC project at a time before all of the final investors have made their investments. In that case, the original investor could make the entire original equity investment with the intention of selling its equity share in the CDE to other investors at a time when the financing could be finalized. [11] Pub. L. No. 108-357 (2004). [12] Pub. L. No. 109-135 (2005). [13] IRS Notice 2006-60, I.R.B. 2006-29. [14] Under the new guidelines, a qualifying business for a targeted low- income population is any corporation (including nonprofit corporations) or partnership that meets one of the following three tests: (1) at least 50 percent of the entity's gross income is derived from sales, rentals, service, or other transactions with individuals who are low- income persons; (2) at least 40 percent of the entity's employees are low-income individuals; or (3) at least 50 percent of the entity is owned by low-income individuals. [15] The original allocation schedule was $1 billion in 2001, $1.5 billion in 2002, $1.5 billion in 2003, $2 billion in 2004, $2 billion in 2005, $3.5 billion in 2006, and $3.5 billion in 2007. [16] Pub. L. No. 109-432 (2006). [17] As of December 2006, only six QEIs had been made into fourth round allocatees. The 2006 NMTC allocations were not announced until the summer of 2006, which may explain the relatively small amount of investment activity into fourth round allocatees at the time of this report. [18] This reflects data available in the CDFI Fund's databases through fiscal year 2005 for awardees. Because of the timing of CDE reporting requirements--CDEs are not required to report data about low-income community investment to the CDFI Fund until 6 months after the end of their fiscal year--it is likely that more NMTC investment has taken place that has not yet been recorded in the CDFI Fund's databases. [19] Unlike tiered NMTC investments, there is no standard term for one investor making a QEI into a CDE. For the purposes of this report, we refer to this type of transaction as a "direct" NMTC investment. [20] Before making an investment in a CDE or in another pass-through entity, investors may set up a partnership as a pass-through entity. [21] In Rev. Rule 2003-20, 2003-1 C.B. 465, the IRS, based on the facts presented in the ruling, approved this method of structuring NMTC investments. [22] Loan-to-value ratio is the relationship, expressed as a percentage, between the amount of a loan and the value of the asset that the loan is being used to finance. In the example above, if 100 percent of the proceeds were reinvested in a CDE as a QEI, the loan-to- value ratio would be 60 percent because a $600,000 loan is being issued to finance a project with a total cost of $1 million. [23] NMTC claimants are a subset of the overall population of NMTC investors. Some investors are pass-through entities designed to pool funds before making an investment in a CDE, but they do not claim the tax credit on tax returns. Additionally, because the CDFI Fund does not track when an investor sells its equity share in a CDE to another investor (and thereby transfers the right to claim the tax credit), the data here are not reflective of all NMTC claimants. The data presented here, unless otherwise noted, originate from the CDFI Fund's databases. [24] Term loans are loans that often only require interest payments until the last day of their term at which time the entire principle amount is due. [25] In most cases, median net assets for businesses and median adjusted gross incomes for individuals would have been more appropriate comparison measures. However, we were unable to use the available data to determine median values for each measure presented here. As a result, we present averages instead of medians. [26] The source of the comparison data for both businesses and individuals is IRS's Statistics of Income division data file for taxpayers that filed tax returns in tax year 2003, the most recent year with available data. [27] The measure of income used for individuals is adjusted gross income from their individual income tax form 1040. Adjusted gross income is a tax paying unit's income after subtracting certain deductions from total income. As a result, when we refer to individual investors, we are referring to tax paying entities--adjusted gross income on the form 1040 could include income from multiple individuals who are living in the same household or married taxpayers. [28] Our survey had a 51 percent response rate. We have used type of entity and size of entity to adjust the data for nonresponse bias. If nonrespondents differ in their responses to survey questions beyond the variables used in our adjustment, our estimates will not reflect this difference. We have assumed that the respondents are a stratified random sample of the population. All percentage estimates from the survey are represented at the 95 percent confidence level. In most cases confidence intervals are only reported where at least one estimate's margin of error is greater than 8 percentage points, plus or minus. Where providing comparable statistics in charts and tables, we have provided all confidence intervals. For additional information about our survey methodology, see app. I. [29] The CDFI Fund collects self-reported data on the expected rate of return for NMTC investments from CDEs that make investments. [30] The CDFI Fund defines fixed assets for businesses as a loan or investment that will be used to pay for any tangible property used in the operation of a business, but is not expected to be consumed or converted into cash in the ordinary course of events. Commonly financed fixed assets include machinery and equipment, furniture and fixtures, and leasehold improvements. [31] The CDFI Fund defines working capital as a loan or investment that will be used to cover any ongoing operating expenses of a business, such as payroll, rent, or utility expenses. [32] Most allocatees are using their qualified equity investments from investors to make loans to qualified businesses, but they can also make investments in other, non-related CDEs. Through fiscal year 2005, over 99 percent of NMTC investment dollars had been made to businesses and less than 1 percent to other CDEs. [33] The CDFI Fund includes a variety of categories under what is considered better rates and terms for financial notes that are issued by CDEs as qualified low-income community investments. In general, the CDFI Fund deems a financial note to have better rates and terms if the CDEs reporting the investment indicate that the rates or terms associated with the investment could not have been offered by the allocatee or otherwise been made available in the marketplace. [34] For the purposes of the 2006 NMTC allocation round, the CDFI Fund defined "areas of higher distress" as areas (1) with poverty rates greater than 30 percent; (2) with median incomes of less than 60 percent of area median income; (3) with unemployment rates at least 1.5 times the national average; (4) that are designated Empowerment Zones, Enterprise Communities, or Renewal Communities; (5) that are U.S. Small Business Administration (SBA)-designated Historically Underutilized Business Zones (HUB Zone), to the extent the investment will support businesses that received HUB Zone certification from the SBA; (6) that are federally designated brownfields redevelopment areas; (7) that are encompassed by a HOPE VI redevelopment plan; (8) that are federally designated as Native American or Alaskan Native areas, Hawaiian Homelands, or redevelopment areas by the appropriate tribal or other authority; (9) that are designated as distressed by the Appalachian Regional Commission or Delta Regional Authority; (10) that are Colonias areas designated by the Department of Housing and Urban Development; (11) that are federally designated medically underserved areas, to the extent the investment will result in the support of health-related services; (12) that are CDFI Fund Hot Zones; (13) that are High Migration Rural counties; (14) that are state or local tax increment finance districts, enterprise zone programs, or other similar state/ local programs targeted toward particularly economically distressed communities; or (15) that are counties for which FEMA has (a) issued a major disaster declaration since July 15, 2005 and (b) made a determination that such county is eligible for both "individual and public assistance." [35] For instance, new investments might be funded by a decrease in dividend payouts for businesses. [36] See app. II for a more thorough description of the steps we took to verify the validity of these baseline statistical results. [37] As described in app. I, the survey population and the statistical analysis population of NMTC investors are not identical. We surveyed NMTC investors that we identified using CDFI Fund data, and in a limited number of cases we surveyed a point of contact at a pass- through entity rather than all of the investors in the pass-through entity. Our statistical analysis population of NMTC investors includes NMTC participants that we identified as credit claimants using both IRS and CDFI Fund data. [38] Potential economic costs are often referred to by economists as efficiency costs or deadweight losses. Efficiency costs result when tax provisions cause individuals or businesses to alter decisions like how much to work, how much to save, what to consume, and where to invest. An exception would be the case where the tax credit is offsetting a market failure, such as a shortage of capital funds available in low- income communities for reasons other than economic returns. Potential benefits include the extent to which a community experiences reductions in poverty and increases in employment opportunities as a result of the program; possible "spillover" benefits to the community may include reductions in crime and improvements in the health status of community residents. [39] This conclusion follows from the statistical evidence alone and does not depend on combining evidence from the survey as was the case for shifted investment for businesses. New investment for individuals is funded through a decrease in consumption (e.g., the amount spent on goods and services). [40] See app. II for more information on the methods we used to develop these statistical models. [41] Our analysis does not address the question of whether NMTC investment by individuals would have taken place by different investors if these particular investors did not make NMTC investments. Our analysis is limited because it only allows us to say that the NMTC investment was new investment by these particular investors. [42] To assess whether the funds would have been used in a more beneficial way in the absence of the program, one would need information on both the financial returns to the alternate use and any positive "spillover" benefits created by NMTC investments such as creating a more skilled workforce. [43] As of November 2006 IRS was able to verify that 61 of the 66 allocatees that received NMTC awards in 2003 had filed tax returns. For those instances where they were still unable to associate a filing, IRS is making an effort to contact the CDE to determine why they have not yet filed a tax return. [44] IRS officials have not developed specific criteria for what "meaningful results" include, but they indicated that they intend to continue conducting NMTC audits until they are comfortable that they have identified any key compliance issues that may arise. [45] GAO, Tax Administration: New Compliance Research Effort Is on Track, but Important Work Remains, GAO-02-769 (Washington, D.C.: June 27, 2002). [46] GAO, Government Auditing Standards, 2003 Revision, GAO-03-673G (June 2003). [47] GAO-04-326. [48] GAO, Tax Credits: Opportunities to Improve Oversight of the Low- Income Housing Program, GAO/GGD/RCED-97-55 (Washington, D.C.: Mar. 28, 1997). [49] If IRS finds a noncompliant CDE, it indicated that it will request an investor list from the CDE to take appropriate action. If the investment was sold after its original issuance, IRS plans to obtain information from the known investors regarding the purchaser of the investments until the total recapture amount is accounted for. [50] Our survey only included tax credit claimants and, in a limited number of cases, a point of contact at a pass-through entity as identified in CDFI Fund data. It did not include lenders participating in NMTC leveraged transactions, which accounts for just over one- quarter of the total amount of QEI. Data are not available on lenders in leveraged transactions. [51] See the Literature Review and Credits section at the end of this appendix for a list of references. [52] For corporate filers, the entire population of returns was drawn for the top two quintiles because there were less than 4,000 total returns in these quintiles. For individual filers, the bottom two quintiles were divided into wage-only and other income groups. Most NMTC claimants had some business income but a few had only wage income, making them harder to distinguish from the "average taxpayer" for whom wage-only income is common. Separating wage-only and other income groups allowed us to minimize the number of returns drawn from the wage- only subset of filers (who we determined to be, in general, less "like" the NMTC claimants). [53] Note that we follow the SCF in excluding tax-preferred retirement accounts, which may cause our wealth estimates to underestimate the wealth of NMTC investors relative to noninvestors as NMTC investors tend to have higher incomes than noninvestors. However, the exclusion's effect, if any, on relative growth rates is not clear. This exclusion is primarily caused by data limitations, as our income tax data only include information for taxable distributions from these accounts (not applicable for most filers) not contributions. For more information on SCF methods, see Arthur B. Kennickell "The Good Shepherd: Sample Design and Control for Wealth Measurement in the Survey of Consumer Finances" (SCF Working Paper, Federal Reserve, Washington, D.C., 2005). [54] Regions were chosen in accordance with U.S. Census Bureau divisions. [55] See Kenneth J. Stewart and Stephen B. Reed, "CPI Research Series Using Current Methods, 1978-98," Monthy Labor Review, vol. 122, no. 6 (1999), for more information and access the data at: [Hyperlink, http://www.bls.gov/cpi/cpiurstx.htm]. [56] We are likely to be understating wealth for households in markets that grew at historically fast rates from 2001 through 2004. We predicted home equity values based on 2000 tax return data and used inflation adjustments to obtain values for 1997 through 1999 and 2001 through 2004. The most likely affect is to bias our results downward because NMTC investors are more likely to have more expensive homes (they are higher income on average than noninvestors, and this factor is associated with higher values of home equity) and experience a greater increase in wealth from the increase in housing values. Consequently, wealth for investors would be underestimated to a larger degree than that of noninvestors and our analysis may underestimate the effect of the NMTC on the wealth of NMTC investors. [57] All dollar amount variables were adjusted using the CPI (research series). The log transformation is used for all variables except for net assets, which are transformed using the inverse hyperbolic sine (to better address negative and zero values). [58] The average treatment effect on the growth in assets for individuals was from 2000 to 2004. For corporations, the period was 2000 to 2003 (we choose 2003 for corporate filers because of the number of filers not in our records for 2004). [59] Initial Hausman tests indicated that fixed effects estimation was more appropriate than random effects estimation. [60] The Y, X and µ variables are time-demeaned data. The inverse hyperbolic sine transformation is used for net assets. [61] Our regression techniques included using the change in the inverse hyperbolic sine of net assets as a dependent variable and median regressions of inverse hyperbolic sine and the change in inverse hyperbolic sine. Our propensity scoring approach was to compare the change in inverse hyperbolic sine of NMTC claimants to their nearest neighbor (closest match) based upon their propensity score using differences in 2000 and 2003 data. [62] A regression model suffers from endogeneity if one or more of the explanatory variables is correlated with the error term (the unexplained portion of the variance in the dependent variable). [63] The household specific effect (i) is fixed over time and differenced out of the equation. [64] An instrumental variables estimator is a method in which another variable that is not correlated with the error term and is (partially) correlated with the endogenous explanatory variable (NMTC participation) is used to predict the endogenous variable in a separate equation. [65] Our analysis of correlations indicated that the presence of the general business credit is not strongly correlated with wealth, as one might expect. This is likely because claiming these credits is relatively rare for individual filers and presence of the credit might be more indicative of a preference for certain types of tax planning or awareness of tax incentives. [66] STATA's panel IV regression models do not include a categorical variable option for the first-stage regression so we are running a linear probability model for the first stage. [67] Demographic information was not included as those variables are constant over the panel. [68] We experimented with difference-in-log specifications to measure growth in our fixed effects models, but the fit of the models was not sufficient to interpret the results. GAO's Mission: The Government Accountability Office, the audit, evaluation and investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through GAO's Web site (www.gao.gov). Each weekday, GAO posts newly released reports, testimony, and correspondence on its Web site. To have GAO e-mail you a list of newly posted products every afternoon, go to www.gao.gov and select "Subscribe to Updates." Order by Mail or Phone: The first copy of each printed report is free. Additional copies are $2 each. A check or money order should be made out to the Superintendent of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a single address are discounted 25 percent. Orders should be sent to: U.S. Government Accountability Office 441 G Street NW, Room LM Washington, D.C. 20548: To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202) 512-6061: To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov Automated answering system: (800) 424-5454 or (202) 512-7470: Congressional Relations: Gloria Jarmon, Managing Director, JarmonG@gao.gov (202) 512-4400 U.S. Government Accountability Office, 441 G Street NW, Room 7125 Washington, D.C. 20548: Public Affairs: Paul Anderson, Managing Director, AndersonP1@gao.gov (202) 512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149 Washington, D.C. 20548: