Popular in Course
verified elite notetaker
Popular in Taxation
This page Document was uploaded by an elite notetaker on Monday December 21, 2015. The Document belongs to a course at a university taught by a professor in Fall. Since its upload, it has received 31 views.
Reviews for Tax-Arbitrage-by-Colleges-and-Universities
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 12/21/15
CONGRESS OF THE UNITED STATES CONGRESSIONAL BUDGET OFFICE STUDY APRIL 2010 Tax Arbitrage by Colleges and Universities i Emormat e A Ii r 8038 Department of the Treasury a 9 quot Internal Revenue Service 4 Part l Reporting Authority t 1 Issuer s name I l L 77W 7 3 Number and street or FRO box if mai te mt eliv d to street a 7 Name of issue 5 City town or post office state and ZIP cede 39 quot tea 39 rage eu an l ggnat t iutl ec tit 3 f V 1 t0 7 Dquot a bet ons iogk k 7 F xepoft mum n few of EL 3 R0 9 te 0 SE Ue SS 1 d to street addre 32 t tteVN 979 USP numb d q H oer of er of OW Pub No 4015 STUDY H m 0 Tax Arbitrage by Colleges and Universities April 2010 The Congress of the United States I Congressional Budget Office Preface Because colleges and universities serve a public purpose advancing higher education and promoting myriad forms of research they enjoy a variety of tax preferences In addition to being exempt from paying federal income taxes institutions of higher learning can accept tax deductible charitable contributions and use tax exempt debt to finance capital expendi tures It is the latter preference that the Congressional Budget Of ce CBO focuses on in this study which was prepared at the request of the Ranking Member of the Senate Finance Committee The law explicitly prohibits the use of tax exempt bond proceeds for the pur chase of investment assets a practice known as tax arbitrage however issuers of tax exempt bonds may use the proceeds for the purchase of operating assets while they simultaneously hold investment assets that provide a higher rate of return To the extent that colleges and universities earn an untaxed return on investments that exceeds the interest they pay on tax exempt debt they are benefiting from a form of indirect tax arbitrage Using data from information returns filed with the Internal Revenue Service by institutions of higher learning and by issuers of tax exempt debt CBO created several measures of tax arbitrage under a broader definition of the term that includes indirect tax arbitrage Over time if legislators were to expand the definition of tax arbitrage nonprofit institutions would most likely respond by reducing their issues of tax exempt debt That response in turn could decrease the cost to the federal government of granting such tax preferences In accordance with CBO s mandate to provide objective nonpartisan analysis the paper makes no recommendations Kristy Piccinini of CBO s Tax Analysis Division wrote the study under the supervision of Frank Sammartino and G Thomas Woodward formerly of CEO Nabeel Alsalam Robert Dennis Mark Hadley and Deborah Lucas provided helpful comments In addition Thomas Pollack of the National Center for Charitable Statistics provided assistance with the data and William Gentry of Williams College Thomas Holtmann of the Joint Committee on Taxation Kim Reuben of the Urban Institute and Dennis Zimmerman of the American Tax Policy Center commented on earlier drafts The assistance of external reviewers implies no responsibility for the final product which rests solely with CBC Loretta Lettner edited the study and Kate Kelly proofread it Maureen Costantino designed the cover and Jeanine Rees prepared the report for publication Lenny Skutnik printed the initial copies Linda Schimmel coordinated the print distribution and Simone Thomas prepared the electronic version for CBO s Web site wwwcbogov Mgabd maxf Douglas W Elmendorf Director April 2010 Contents Summary and Introduction Tax Preferences for Higher Education TaxExempt Bonds and Tax Arbitrage The Use of Tax Exempt Bonds Advantages and Disadvantages of Tax Exempt Bonds Direct and Indirect Tax Arbitrage Approaches to Measuring the Extent of Indirect Tax Arbitrage Investment Assets and Tax Exempt Debt Possible Approaches to Expanding the Definition of Tax Arbitrage Measuring the Volume of Arbitrage Bonds Under a Broader Definition of the Term Estimated Volume of Arbitrage Bonds Under a Broader De nition of the Term Estimated Amount of Arbitrage Debt The Distribution of Arbitrage Debt Comparison with Arbitrage Debt for Nonprofit Hospitals Institutional Response to an Expanded Definition of Tax Arbitrage Appendix Alternative Calculations of Tax Arbitrage as Practiced by Colleges and Universities quotAK10000 l 10 12 12 13 14 15 16 17 VI TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES Tables 1 Uses of Proceeds from Tax Exempt Bonds Issued on Behalf of Colleges and Universities 2003 2 Selected Assets and Liabilities Held by Colleges and Universities 2003 3 Tax Exempt Debt Classi ed as Earning Pro ts from Arbitrage in 2003 Under a Broader Definition of the Term 4 Colleges and Universities Conducting Tax Arbitrage in 2003 Under a Broader Definition of the Term Figures 1 The Value of New Issues of Tax Exempt Bonds 1990 to 2007 2 The Distribution of Investment Assets and Outstanding Tax Exempt Bonds Held by 251 Colleges and Universities in 2003 3 The Distribution of Tax Exempt Debt Held by 251 Colleges and Universities in 2003 That Would Be Classi ed as Earning Pro ts from Arbitrage Under a Broader De nition of the Term Box 1 Estimating the EXtent of Indirect Arbitrage Practiced by Colleges and Universities 13 14 10 15 Tax Arbitrage by Colleges and Universities Summary and Introduction Colleges and universities enjoy a variety of federal tax preferences that are designed to support a broader public purpose the advancement of higher education and research Not only are institutions of higher learning exempt from paying federal income taxes they also are eligible to receive tax deductible charitable contributions and allowed to use tax exempt debt to finance capital expenditures This Congressional Budget Of ce CBO study focuses on one of those tax advantages the ability of colleges and universities to borrow funds by issuing tax exempt debt According to the staff of the Joint Committee on Taxation J CT the cost of allowing institutions of higher learning to borrow using such debt measured in terms of the revenues that could have been collected if those institutions had borrowed using taxable debt will be about 55 billion in 2010 The use of proceeds from lower cost tax exempt bonds to directly finance the purchase of higher yield securities a practice known as tax arbitrage is prohibited by law Nevertheless the law as currently implemented allows many colleges and universities to use tax exempt debt to finance investments in operating assets buildings and equipment while at the same time they hold investment assets that earn a higher return Investment assets are publicly traded and privately held securities as well as land or buildings held for investment purposes To the extent that colleges and universities can earn untaxed returns on investments that are higher than the interest they pay on tax exempt debt they are bene ting from a form of indirect tax arbitrage Rules in the Internal Revenue Code IRC and regula tions established by the Department of the Treasury limit tax arbitrage by restricting the yield on any investments held by the bond issuer that are deemed to be directly related to the tax exempt bond issue for example an asset pledged as collateral1 Other investment assets are not yield restricted even though they contribute indi rectly to securing the bonds and are considered by rating agencies when rating the tax exempt debt A broader definition of tax arbitrage would include most or all investment assets held by an institution borrowing with tax exempt debt Using data from information returns led with the Internal Revenue Service IRS by institutions of higher learning and by issuers of tax exempt debt CBO devel oped measures of tax arbitrage as practiced by colleges and universities under a broader definition of the term that encompasses both direct and indirect tax arbitrage Under one such de nition nearly all of the tax exempt bonds that 251 institutions issued in 2003 would be clas si ed as earning pro ts from tax arbitrage If some invest ment assets were set aside in a reserve which would be excluded from the arbitrage measure under an alternative expanded de nition the amount of debt earning returns from arbitrage would be lower even so about 75 percent of bonds issued in 2003 would still be classi ed as earn ing arbitrage pro ts under that expanded de nition By either measure the amount of debt issued by colleges and universities that earns arbitrage pro t would be consider ably larger than that issued by nonpro t hospitals which was the subject of a previous CBO study on broadening the de nition of tax arbitrage2 Over time if legislators were to expand the de nition of tax arbitrage and thereby eliminate some of the bene ts of tax exempt nancing 1 Internal Revenue Code 26 USC 148 b 3A The terms debt and bond are used interchangeably to refer to debt with maturities in excess of a year The dollar gures for such debt cited in this analysis also include any leasing arrangements that are tax exempt 2 Congressional Budget Of ce Nonpro t Hospitals and Tax Arbitrage letter to the Honorable William M Thomas December 6 2006 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES nonpro t institutions would probably respond by reduc ing the issuance of tax exempt debt That response in turn would decrease the cost to the federal government of the tax preference Tax Preferences for Higher Education Institutions of higher learning both public and private benefit from several types of preferential tax treatment Like other nonprofit organizations defined in section 501c 3 of the IRC nonprofit private schools are exempt from the federal income tax eligible to receive charitable contributions that donors may deduct from their taxable income and allowed to use tax exempt debt to finance capital expenditures3 As state or local govern ment entities public colleges and universities receive broadly similar tax preferences they are exempt from fed eral income taxation eligible for donations that are tax deductible and may have access to tax exempt debt Although there are no estimates of the cost to the federal government of exempting contributions made specifically to colleges and universities the deduction of charitable contributions to educational institutions at all levels is expected to cost about 66 billion in forgone tax reve nues in 2010 charitable contributions to colleges and universities account for about 70 percent of all contribu tions to educational institutions4 JCT estimates that allowing institutions of higher learning to borrow using tax exempt debt will cost the federal government in the form of forgone tax revenues about 55 billion in 2010 As is the case with other nonprofit organizations colleges and universities receive preferential tax treatment because they are viewed as serving a public purpose Institutions of higher learning perform two activities that are typically considered to serve the needs of society providing educa tion and conducting research Education is associated with a wide range of favorable outcomes Investment in 3 Section 501c3 of the Internal Revenue Code defines a quali fied nonprofit as any entity organized and operated exclusively for religious charitable scientific testing for public safety literary or educational purposes or to foster national or international amateur sports competition or for the prevention of cruelty to children or animals A nonprofit may not engage in political activity and none of its earnings may benefit any private share holder or individual 4 Joint Committee on Taxation Estimates of Federal Tax Expendi tares for Fiscal Years 2008 2012 JCS 2 08 October 31 2008 p 53 human capital through education confers considerable private benefits on an individual in the form of higher income and better health6 Education probably also yields benefits for the community as a whole including a more productive workforce which leads to faster economic growth as well as lower crime a more informed elector ate and increased social mobility7 How much education an individual prefers to invest in depends solely on the private benefit he or she might expect from that invest ment in the absence of government intervention that decision will yield fewer public benefits than is socially desirable8 Some colleges and universities also perform research that may have large spillover effects that benefit the rest of the economy9 Although businesses make substantial invest ments in research and development private investors cannot retain all of the benefits from that spending because the knowledge produced by such research can be used by others As is the case with individuals who must decide how much to invest in their own education the private sector chooses the amount of research and devel opment it is willing to fund on the basis of private rather than social benefit 5 Estimates of tax expenditures are not intended to capture all of the ways in which taxpayers might respond to a change in law In par ticular the estimate discussed here assumes that if investors did not hold tax exempt bonds they would hold taxable bonds instead In one study researchers examined some of the other ways in which investors might change their portfolios in response to limits on tax exempt bonds That study found that because tax able bonds are one of the most heavily taxed types of asset inves tors would probably seek alternatives that are less heavily taxed and therefore the revenue loss to the federal government would probably be smaller than the tax expenditure cost discussed here See James Poterba and Arturo Verdugo Portfolio Substitution antl toe Revenue Cost ofExempting State and Local Government Interest Payments om Federal Income Tax Working Paper No 14439 Cambridge Mass National Bureau of Economic Research October 2008 available at wwwnberorgpapersw14439 6 See David Card The Causal Effect of Education on Earnings in Orley Ashenfelter and David Card eds Handbook ofLaoor Economics vol 3 Amsterdam Elsevier Press 1999 pp 1801 1863 Researchers in another study discuss the evidence for a pos itive relationship between education and health outcomes paying particular attention to the mechanisms through which education may lead to better health See David M Cutler and Adriana Lleras Muney Education and Health Evaluating Theories and Evi dence in Robert F Schoeni and others eds Makinglmericans Healtoier Social and Economic Policy as Healto Policy New York Russell Sage Foundation January 2008 Granting favorable tax treatment to postsecondary educa tional institutions is just one way in which policymakers may be able to increase investment in human capital and research Other ways that the federal government cur rently subsidizes educational institutions include direct grants to states and localities for elementary and second ary education tax preferences for private institutions that provide elementary and secondary education direct grants to schools for research and subsidies and loan programs for individuals pursuing undergraduate and graduate degrees TaxExempt Bonds and Tax Arbitrage State and local governments use tax exempt bonds to finance their own capital projects and to provide the means for other entities including nonprofit and state supported colleges and universities to use tax exempt 7 See Enrico Moretti Estimating the Social Return to Higher Edu cation Evidence from Longitudinal and Repeated Cross Sectional Data journal ofEconometrics vol 121 no 1 2 July August 2004 pp 175 212 In his analysis Moretti found that college education creates positive spillovers in productivity and wages See also Eric Hanushek and Ludger Woessmann Do Better Scoools Leaol to More Growto Cognitive Skills Economic Outcomes and Causation Working Paper No 14633 Cambridge Mass National Bureau of Economic Research January 2009 In their analysis the authors found empirical evidence of a causal relation ship between educational attainment and growth rates across countries Other research has found that educational attainment is associated with a decreased likelihood of incarceration or arrest See Lance Lochner and Enrico Moretti The Effect of Education on Crime Evidence from Prison Inmates Arrests and Self Reports American Economic Review vol 94 no 1 March 2004 pp 155 189 Still other research suggests a positive correlation between educational attainment and the likelihood of electoral participation See Kevin Milligan Enrico Moretti and Philip Oreopoulos Does Education Improve Citizenship Evidence from the United States and the United Kingdom journal of Pub lic Economics vol 88 no 9 10 August 2004 pp 1667 1695 For a discussion of the relationship between postsecondary educa tion and social mobility see Robert Haveman and Timothy Smeeding The Role of Higher Education in Social Mobility Future ofCoilolren Opportunity in America vol 16 no 2 Fall 2006 pp 125 150 8 Individuals who face financial constraints may invest in less education than is either privately or socially desirable Federal student loan programs are one way to reduce the impact of such constraints See Congressional Budget Office Costs and Policy Options for Feoleral Stuolent Loan Programs March 2010 9 For a more detailed discussion of federal subsidies for research and development see Congressional Budget Office Feoleral Support for Researco and Development June 2007 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES debt The tax code contains provisions that are designed to prevent that tax preference from becoming an unlim ited subsidy for all types of spending and to restrict its use to financing capital investment in operating assets such as the construction or renovation of buildings and the purchase or repair of equipment Tax arbitrage the use of proceeds from lower cost tax exempt bonds to finance the purchase of higher yield securities is specifically prohibited both by the IRC and by Treasury regula tions10 However in most situations the law does not prevent tax exempt borrowers from engaging in what is essentially indirect tax arbitrage Indirect tax arbitrage occurs when a borrower with tax exempt status earns interest on investment assets not directly financed with bond proceeds that exceeds the interest cost incurred from contemporaneous tax exempt borrowing A borrower could sell those assets to finance the capital expenditure instead of borrowing with tax exempt debt Holding those assets while borrowing on a tax exempt basis is in effect equivalent to using tax exempt proceeds to invest in those higher yielding securities The Use of TaxExempt Bonds About 290 billion in tax exempt bonds was issued in 2007 the most recent year for which aggregate data are available up from about 100 billion in 1990 see Figure 1 About 70 percent of those were govern mental bonds which are typically issued by state and local governments for public projects such as the con struction of highways or public schools The payment of interest on those obligations is generally funded through tax revenues The remaining tax exempt bonds were qualified private activity bonds tax exempt bonds issued by state and local governments on behalf of certain private entities or for designated activities11 Eligible activities include financing student loans or mortgages for owner occupied 10 Internal Revenue Code 26 USC 148 Treas Reg 26 CFR 1148 0 11 Such bonds can be issued on behalf of a private entity if more than 10 percent of the proceeds is used for any private business purpose and if more than 10 percent of the payment of principal or interest is secured by an interest in property used for a private business purpose or is derived from payments for property used for a private business purpose Private activity bonds are taxable unless they are issued for a qualified purpose or entity 3 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES Figure 1 The Value of New Issues of TaxExempt Bonds 1990 to 2007 Billions of dollars 300 I PrivateActivity Issues Nonhospital 501c3 Entities 250 D All Other PrivateActivity Issues 200 I Governmental Issues I I I I 150 I 100 I l 50 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Source Congressional Budget Office Notes Privateactivity bonds are issued on behalf of private entities by state and local governments Unless they are issued for specific taxexempt activities or entities interest paid on such bonds is taxable Activities that are eligible for taxexempt status include the financing of student loans or mortgages for owneroccupied housing Entities that qualify for taxexempt status include nonprofit hospitals schools and other qualified 501c3 organizations Aside from nonprofit hospitals the Internal Revenue Service does not separate out data on bond issues for any other type of 501c3 organization The category quotall other privateactivity issues consists primarily of bonds issued on behalf of hospitals mortgage bonds and residential rental bonds Governmental bonds are typically issued by state and local governments for public projects such as the construction of schools or highways housing eligible entities include nonpro t hospitals nonpro t schools and other quali ed 501c 3 organi zations The interest on tax exempt private activity bonds is typically paid with revenue generated by the project that was nanced with the bond proceeds rather than by state and local taxpayers The volume of most eligible types of tax exempt private activity bonds that can be issued in a given year by a given state is limited by the Internal Revenue Code as amended in 1986 but those caps do not apply to bonds issued on behalf of quali ed 501c 3 organizations The only cap that applied to tax exempt bonds issued on behalf of quali ed 501c3 organizations previously 150 million for the bene t of any single organization was lifted in 1997 New issues of tax exempt private activity bonds totaled 87 billion in 2007 Of that amount about 27 billion was issued on behalf of quali ed nonhospital 501c 3 organizations and 17 billion was issued on behalf of quali ed 501c 3 hospitals The IRS does not separate out data on bond issues for any other type of 501c3 organization Of the remaining tax exempt private activity bonds quali ed mortgage bonds and quali ed residential rental bonds were the largest categories by vol ume 14 billion of the former and 7 billion of the latter were issued that year To estimate the amount of new bonds issued speci cally for colleges and universities CBO analyzed all informa tion returns submitted to the IRS for bonds issued on behalf of 501c 3 organizations in 200312 That year bonds issued on behalf of institutions of higher learning accounted for just under 6 billion of the 14 billion in new issues for nonhospital 501c 3 organizations Including hospitals total issues for all 501c3 organiza tions were about 24 billion that year while all tax exempt private activity issues totaled 46 billion 12 Although data from more recent years are now available the most current data available when CBO undertook this analysis were for 2003 Table 1 Uses of Proceeds from TaxExempt Bonds Issued on Behalf of Colleges and Universities 2003 Number of Percentage of Issues Issues Construction andor Expansion of Buildings Academic buildings 42 40 Residence halls 34 32 Student centers 8 8 Athletic facilities 11 11 Equipment 10 10 MaintenanceSafety 45 43 Total 105 na Source Congressional Budget Office based on data provided by issuing authorities in nine states Notes The number of issues in the various categories adds to more than 105 and the percentage of issues in each category adds to more than 100 percent because many projects span multiple categories na not applicable Since those data were collected the market for tax exempt bonds issued by institutions of higher learning and the value of the assets that those institutions hold have been greatly affected by the nancial crisis that began in 2007 Interest rates for tax exempt debt rose sharply during that period and issues by colleges and universities have probably declined since the crisis began However there are signs that the pressures created by the nancial crisis are beginning to ease In particular the difference between interest rates on tax exempt debt and those on short term Treasury bonds a standard measure of the risk premium that investors require in order to hold the bonds has fallen The liquidity problems that some borrowers face may decrease the use of certain types of short term debt but that decrease seems unlikely to have a permanent effect on the availability of credit for long term capital needs Because of widespread declines in asset prices educational endowments have fallen in value considerably from their peak but they retain some of the bene t from previous years of growth This analy sis is intended to capture the effect of broadening the TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES de nition of tax arbitrage in the long term rather than the effects of the recent disruptions in nancial markets CBO also collected data on projects that were nanced by tax exempt bonds issued on behalf of institutions of higher learning from issuing authorities in nine states in 2003 covering 23 billion in issues about 40 percent of all issues in that year The most common use of proceeds from tax exempt bonds issued on behalf of col leges and universities in 2003 was for maintenance proj ects such as improved heating and cooling systems and safety enhancements such as sprinkler systems About 43 percent of all bond issues involved such projects Pro ceeds from a single issue may be used for projects in mul tiple categories see Table 1 About the same number of bond issues 40 percent involved the construction and or expansion of academic buildings the next most com mon use nearly 30 percent was for the construction and or expansion of residence halls The use of bond proceeds for athletic facilities student centers or the pur chase of equipment was considerably less common Advantages and Disadvantages of TaxExempt Bonds Compared with other ways the federal government could choose to subsidize colleges and universities tax exempt bonds have both advantages and disadvantages Because nonpro t institutions of higher learning are exempt from the income tax any further subsidy through the tax code must be indirect which leads to one disadvantage The tax exempt bond subsidy is routed through investors who are willing to accept a rate of return on a tax exempt bond that is lower than the return on a taxable bond by the amount they would have to pay in taxes on income from the taxable bond As long as the supply of tax exempt bonds exceeds the demand from taxpayers in the highest income tax bracket the market interest rate on such bonds needs to fall below the rate on taxable bonds only by enough to induce taxpayers in a lower tax bracket to also hold the bonds that rate is higher than what would be necessary to attract investors in the highest tax bracket Therefore investors in the highest bracket receive the interest tax free at their higher marginal tax rate retaining some of the value of the subsidy rather than passing it on to the issuer of the bonds Another disadvantage is that in contrast with federal spending programs tax expenditures including for gone revenues on tax exempt bonds are not explicitly 5 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES identi ed in the budget13 Also unlike discretionary spending programs they are not governed by the annual appropriation process Thus the federal government does directly control the issuance of tax exempt bonds which is determined by state and local issuers in accordance with federal rules on the total volume and type of issue Though state and local governments may be better equipped to identify beneficial capital investments by their local institutions delegating decisionmaking power away from the federal government increases the probability that bond issues will be evaluated on the basis of their benefit to the locality rather than to the federal taxpayers who finance the subsidy One advantage to using tax exempt bonds as the means for offering a subsidy is that they provide a standard framework through which educational institutions can access capital markets Access to the tax exempt bond market may increase the availability of bond financing for some educational institutions although the schools most likely to be affected by the expanded definition of arbi trage discussed in this study have investment grade credit ratings and therefore probably would be able to obtain funding in the taxable bond market The use of tax exempt bonds also affects the allocation of resources The lower cost of financing for projects funded by tax exempt bonds diverts resources toward those proj ects and away from other activities Whether that is an advantage or a disadvantage depends in part on whether the subsidized investment would have been undertaken even in the absence of the subsidy On the one hand if colleges and universities use tax exempt financing for projects that they would complete even without the sub sidy resources are just reallocated from taxpayers to the schools with no additional social benefit On the other hand if the subsidy finances capital projects that would not otherwise have been undertaken and that create a social benefit in addition to the benefit to the institution it could improve the nation s overall welfare 13 The Administration provides estimates of tax expenditures in the Analytical Perspectives volume of the budget See Office of Man agement and Budget Analytical Perspectives Budget oft7e US Government Fiscal Year 2011 Chapter 16 The Joint Committee on Taxation also reports annually on tax expenditures See Joint Committee on Taxation Estimates of Federal Tax Expenditures for Fiscal Years 2009 2013 JCS 1 10 January 1 1 2010 Direct and Indirect Tax Arbitrage Because the purchasers of tax exempt bonds do not pay income tax on the interest those bonds earn they are willing to accept a lower rate of interest than they would otherwise earn on taxable bonds of comparable risk and maturity That yield differential presents an opportunity for some issuers of tax exempt debt to engage in tax arbitrage borrowing with tax exempt debt and invest ing the proceeds in higher yielding taxable assets Those who meet the criteria for borrowing using tax exempt bonds whether quali ed 501c 3 organizations or state and local government entities have an added incentive because they do not pay tax on their net income regardless of whether it is from an operating surplus the excess of revenue over cost or from invest ment income The higher return on those taxable assets not only finances the lower interest cost of the tax exempt debt but also provides untaxed earnings to be used for other purposes To restrict such activity the tax code specifies that arbitrage bonds are not tax exempt Section 148 of the IRC defines an arbitrage bond as any bond whose proceeds are reasonably expected to be used directly or indirectly to acquire higher yielding investment assets or to replace funds which were used directly or indirectly to acquire higher yielding securities 14 The tax code has provisions that prevent the direct diversion of bond pro ceeds away from investment in physical capital to the earning of investment income In general those provi sions allow earnings from tax arbitrage only for tempo rary periods before the proceeds are needed to fund the project for which they were designated or for specific types of investments such as reserve funds Such funds typically contain a portion of the proceeds from a bond issue that is set aside to pay debt service in case the expected sources of funds for that purpose are not avail able Outside of those limited exceptions however any earnings from tax arbitrage must be rebated to the Treasury The Treasury regulates arbitrage using a replacement proceeds rule that requires the yield to be restricted on any investment assets or other amounts that have a connection nexus to a tax exempt bond issue that is sufficiently direct for one to conclude that in the absence of proceeds from tax exempt borrowing the assets or 14 Internal Revenue Code 26 USC 148a amounts would have been used to nance the project15 For example if a school uses securities as collateral for its debt service obligation on a tax exempt bond the securities are treated as replacement proceeds subject to yield restriction The Treasury restricts the yield on investment assets in one of two ways Under one method the borrower may be required to return to the federal government any excess yields earned on an amount of assets that is equal to the value of the tax exempt bonds issued16 Excess yields are defined as earnings that exceed the interest paid on the bond Under a second method the borrower may be required to sell an amount of assets that is equal in value to the bonds issued and to invest that money in a specially designed Treasury debt instru ment that earns a discounted return to offset the federal government s implicit contribution to the return on the tax exempt bonds Those provisions do not eliminate all opportunities for tax arbitrage however Because financial statements typi cally do not report the use of particular assets as collateral the replacement proceeds rule is difficult to enforce In addition if assets are not specifically pledged to pay the debt service on a tax exempt bond or if the assets have no other direct connection to the bonds the arbitrage restrictions do not apply However it is widely recognized that assets and their earnings can be used to pay the inter est on debt or to cover other expenses to free up funds for interest payments regardless of whether they are directly pledged to do so Such use of higher yielding assets to finance tax exempt debt constitutes indirect tax arbitrage The limited scope of the tax code s restrictions on arbi trage is not unique to colleges and universities It applies equally to other nonprofit institutions such as nonprofit hospitals which have sizable financial assets For all such institutions the current tax arbitrage rules ensure that a bond issue is associated with the acquisition of new capi tal and they reduce its cost Nonetheless a change in the rules that broadened the definition of tax arbitrage would 15 Treas Reg 26 CFR 1148 1c 16 Such payments must be made every five years during the life of the issue with the first payment made no later than five years after the issue date of a bond If the computation in later years shows no arbitrage profit because the yield on restricted assets has declined previous payments are refunded to the issuer 17 It is also standard practice for rating agencies to base credit ratings for a particular debt issue on all available assets not just on those directly pledged to that debt issue TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES identify bonds earning arbitrage profits on the basis of the total assets that were implicitly available as collateral rather than requiring a direct relationship between pro ceeds from tax exempt bonds and investment assets explicitly pledged as collateral That expanded definition which would encompass indirect tax arbitrage would decrease the value of the federal subsidy that is currently available to institutions of higher learning through tax exempt bond issues and reduce the net cost of the tax exemption to the federal government Approaches to Measuring the Extent of Indirect Tax Arbitrage Determining the degree to which colleges and universities benefit from the practice of indirect tax arbitrage requires data on the volume of new and outstanding issues of tax exempt bonds and on the value of investment assets held by those institutions of higher learning CBO collected data on assets and liabilities from IRS information returns specifically Forms 990 and 8038 adjusting the data to account both for the misreporting of tax exempt liabilities and for the presence of assets held by other organizations for the use of colleges and universities see Box 1 To estimate the extent to which indirect tax arbi trage occurs CBO compared an institution s outstanding bond issues bonds that have not been completely retired or new bond issues with the value of its existing invest ment assets If an institution held assets that were greater in value than its holdings of either outstanding or new bond issues those bond issues were classified as earning returns from tax arbitrage Presumably the school chose to use tax exempt debt to finance capital projects rather than selling investment assets because it could earn a rate of return on those assets that was higher than the interest it was obligated to pay on the bonds If the dollar value of investment assets was less than that of outstanding or new bond issues only the portion of capital spending that could have been financed with the assets was considered to be earning arbitrage profit CBO also calculated esti mates of tax arbitrage allowing some investment assets to be set aside in a reserve that would be exempt from the broader definition of arbitrage Investment Assets and TaxExempt Debt Colleges and universities in aggregate hold investment assets that are significantly higher in value than the stock of outstanding tax exempt bonds although the distribu tion of both is highly skewed That relationship holds both for amounts reported on IRS Form 990 returns and 7 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES Box 1 Estimating the Extent of Indirect Arbitrage Practiced by Colleges and Universities As nonpro t entities institutions of higher learning must le Form 990 information returns with the Internal Revenue Service IRS on an annual basis that provide data from their balance sheets on reve nues expenses liabilities and assets The Congressio nal Budget Of ce CBO used information from a sample of those returns which are available from the IRS to calculate the investment assets and out standing stock of tax exempt debt weighted to be representative of all schools that led Form 990 returns in 2003 CEO also used Form 990 returns from the IRS sample supplemented by additional Form 990 returns from the National Center for Charitable Statistics NCCS at the Urban Institute to calculate the investment assets and outstanding stock of tax exempt bonds for each institution that issued new tax exempt debt in 2003 To identify the latter group CBO used information from Form 8038 information returns which all issuers of tax exempt debt are required to le annually Those returns were then used to estimate the volume of tax exempt bor rowing by institutions of higher learning in 2003 CEO found that Form 8038 returns were led for 324 bond issues made on behalf of colleges and uni versities Of those 58 were classi ed as refunding bonds that re nanced debt already issued in previous years Proceeds from the remaining 266 bond issues which amounted to about 6 billion were distrib uted to 274 institutions some bonds were issued on behalf of more than one organization Of those 274 institutions CBO was able to identify and match 251 with information culled from Form 990 returns obtained either from the IRS sample or from the NCCS Those 251 institutions received 57 bil lion or 94 percent of all proceeds from tax exempt bonds issued on behalf of higher education in 2003 One dif culty with measuring tax exempt debt is that when ling Form 990 returns institutions often misallocate balance sheet information on tax exempt liabilities In an earlier study of tax exempt bonds issued on behalf of nonpro t hospitals CBO found that the majority of tax exempt bond liabilities were misreported as mortgages and other notes payable on Form 990 returns1 Issuers of tax exempt bonds for educational institutions con rmed that such mis reporting is present in returns for colleges and univer sities as well In many cases schools consider the pro ceeds of an issue as a loan from the issuing authority rather than as a tax exempt liability CBO therefore adjusted the stock of tax exempt bonds reported on Form 990 returns to account for that misreporting using the factors estimated in CBO s earlier study of hospitals2 In some cases the value of investment assets reported on Form 990 returns for institutions listed as the receiver of bond proceeds underestimated the true value of investment assets available to the institution For public universities particularly those with large endowments the majority of investment assets are held not by the university but by a foundation dedi cated to the support of the institution and separately incorporated as a public charity That separation allows state supported institutions to exercise greater control over their endowment assets Of the group of 251 borrowers CBO was able to identify 27 insti tutions that held assets in other 501c3 organiza tions CBO adjusted their stock of investment assets and tax exempt liabilities to reflect the assets and liabilities of those related organizations 1 See Congressional Budget Of ce Nonpro t Hospitals and Tax Arbitrage letter to the Honorable William M Thomas December 6 2006 In its analysis CBO found that for those institutions reporting no tax exempt liabilities 996 percent of their reported mortgage liabilities were tax exempt liabilities When an institution reported both tax exempt liabilities and mortgages on average 84 percent of mortgage liability was actually tax exempt bond liability 2 When an institution reported no tax exempt liabilities CBO reclassi ed all mortgage liabilities as tax exempt liabilities When an institution reported both tax exempt liabilities and mortgages CBO reclassi ed 84 percent of mortgage liability as tax exempt bond liability Such adjustment results are presented in the appendix Table 2 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES Selected Assets and Liabilities Held by Colleges and Universities 2003 Millions of dollars Institutions that All Institutions Borrowed in 2003 Total Median Total Median Net Investment Assets Unadjusted 262151 151053 37 Net Investment Assets Adjusted a a 152324 39 Outstanding TaxExempt Debt Reported 44326 0 16901 10 Mortgage Debt Reported 17524 1 5935 3 Outstanding TaxExempt Debt Adjusted 60442 2 22199 22 Memorandum New Bond Issues na na 5703 10 Number of Institutions 913 na 251 na Source Congressional Budget Office based on data from information returns Forms 8038 and 990 filed with the Internal Revenue Service Notes Investment assets include publicly and privately held securities as well as land and buildings held for investment purposes net of accumulated depreciation Taxexempt debt is frequently misreported as quotmortgages and other notes payable If no taxexempt debt is reported the adjusted stock of taxexempt debt includes all mortgages reported by the institution If both taxexempt debt and mortgages are reported the adjusted stock of taxexempt debt includes 84 percent of mortgages reported The value of investment assets reported on a school s Form 990 return does not include any assets in related organizations dedicated to the support of the institution and separately incorporated as a public charity Of the group of 251 borrowers CBO was able to identify 27 institutions that held assets in other 501c3 organizations and adjusted their stock of investment assets and taxexempt liabilities to reflect the assets and liabilities of those related organizations na not applicable a CEO identified related organizations just for new borrowers in 2003 for those amounts adjusted to correct for misreporting see BOX 1 Colleges and universities reported about 260 billion in total investment assets and about 45 billion in liabilities for tax exempt bonds in 2003 see Table 2 After adjusting for misreporting CBO estimated that the outstanding stock of tax exempt debt was about 60 billion The median amount of invest ment assets reported on Form 990 returns for 2003 was about 4 million that is half of the institutions had more than 4 million in assets and half had less and the median amount of tax exempt debt was zero With adjustments for misreporting the median stock of tax exempt debt increased to about 2 million The subset of schools that borrowed using tax exempt debt in 2003 also had total investment assets that far exceeded tax exempt liabilities Those schools total investment assets at about 150 billion were nine times larger than the total reported stock of outstanding tax exempt bonds and seven times larger than the total adjusted stock of outstanding tax exempt bonds The median school that borrowed in 2003 had an estimated 39 million in investment assets The median stock of tax exempt liabilities was about 22 million for schools that borrowed in 2003 about 40 percent less than the median amount of investment assets Within the group of 2003 borrowers the distributions of investment assets and tax exempt bonds were highly skewed see Figure 2 The top 10 schools that borrowed in 2003 ranked by investment assets made up about 4 percent of the sample but held almost 75 percent of the total amount of investment assets held by the entire group The top 50 schools that borrowed in 2003 ranked by investment assets made up about 20 percent of the sample but held about 95 percent of total investment assets The group of institutions with the largest share of investment assets had also issued a substantial share of the tax exempt bonds but the distribution of the bond issu ances was less skewed The top 10 schools that borrowed 9 1O TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES Figure 2 The Distribution of Investment Assets and Outstanding TaxExempt Bonds Held by 251 Colleges and Universities in 2003 Percent 80 70 I Investment Assets Adjusted 60 Outstanding Stock of 50 TaxExempt Bonds 40 Adjusted 30 20 10 1 to 10 11 to 50 51 to 100 101 to 150 151 to 200 201 to 251 Institutions Ranked According to Reported Investment Assets Source Congressional Budget Office Notes In an earlier study of taxexemptbond issues made on behalf of nonprofit hospitals CBO found that the majority of taxexempt bond liabilities were misreported as quotmortgages and other notes payable on Form 990 returns see Congressional Budget Office Nonpro t Hospitals and Tax Arbitrage letter to the Honorable William M Thomas December 6 2006 Issuers of taxexempt bonds for educational institutions confirmed that such misreporting is present in returns for colleges and universities as well CBO therefore adjusted the stock of taxexempt bonds reported on Form 990 returns to account for that misreporting using the factors estimated in the earlier study Some public schools hold assets in separate 501c3 foundations dedicated to the support of the institution CBO adjusted the amount of investment assets held by 27 institutions to reflect such assets held by foundations in 2003 accounted for about 40 percent of the out standing value of such bonds and the top 50 accounted for almost 75 percent of the outstanding value The majority of outstanding tax exempt bonds were held by schools with substantial investment assets which would probably allow them to borrow even if tax exempt borrowing was not an option This suggests that as cur rently implemented the subsidy is not used primarily to ease access to nancial markets for schools that would otherwise have dif culty undertaking capital projects Possible Approaches to Expanding the De nition of Tax Arbitrage Any speci cation of the investment assets that would be covered under an expanded de nition of tax arbitrage should account for the legitimate role that such assets play in the operation of colleges and universities Those institutions accumulate investment assets for a variety of reasons to earn income to ful ll the purposes that qualify them for tax exempt status to protect against uncer tainty to obtain a stronger credit rating to enhance their reputation and to honor gift restrictions Some of those reasons suggest that certain investment assets should not be counted when measuring earnings from tax arbitrage The need to maintain an operating reserve is one ratio nale for allowing an institution to hold some investment assets while issuing tax exempt bonds One possible rea son for accumulating substantial operating reserves is to maintain spending levels as income from endowments and other sources of revenue uctuate but in practice most schools follow self imposed spending rules that limit such smoothing and mandate spending reduc tions when income falls Another possible reason is that rating agencies offer higher credit ratings as the ratio of expendable nancial resources which include operating reserves to yearly expenses increases and those higher ratings lead to lower costs for borrowing18 A comparison of investment assets and annual expenses shows that by rating agencies standards many schools that borrowed in 2003 appear to have substantial operating reserves For instance the median ratio of investment assets to annual expenses was 1 for schools that borrowed in 2003 that is investment assets were equal to about a year s worth of expenses That ratio ranged from a high of 895 to zero Some investment assets held by colleges and universities are subject to restrictions by the donor Because educa tional institutions cannot use such assets for purposes other than those specified by the donor it could be argued that those assets should not be included when measuring earnings from tax arbitrage It might be possi ble however to implement a broader measure of tax arbi trage without forcing schools to violate most of those restrictions For instance if a donor restricted a gift to an academic institution s endowment that is specified that the gift be used to generate future earnings rather than to help pay for current operating expenses the require ment under a broader measure of arbitrage that earnings on the gift be rebated if they exceeded the interest paid on a tax exempt bond would not necessarily violate the donor s restriction The gift itself would not be used to purchase the asset financed by the bonds only the earn ings on the gift would be affected19 In cases where the donor directed that a gift but not the earnings on that gift be used for a specific purpose similar reasoning would apply Even in cases where the donor directed that both the gift and its earnings be used for a specific pur pose many restricted purposes research certain types of academic support athletics would be consistent with using earnings on that gift to finance the construction of academic and athletic buildings laboratory facilities and libraries20 Another consideration is that exempting restricted gifts from an expanded definition of tax arbitrage would strengthen the incentive for schools to pursue restricted gifts Colleges and universities frequently cite restrictions 18 According to Moody s Investors Service the median ratio of unrestricted net resources to expenses was 09 in 2006 See Private College and University Medians 2007 New York Moody s Investors Service May 2007 19 According to a survey conducted by the National Association of College and University Business Of cers about 60 percent of the assets in college and university endowments in 2003 were restricted to income generating purposes TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES 1 1 on gifts as one reason they require federal subsidies for other types of spending that are less preferred by donors or to justify their tax exempt endowments freedom from federally mandated spending requirements such as those that apply to charitable foundations Currently schools must weigh the support that additional restricted gifts lend to those arguments against the fact that unrestricted gifts allow the institution more exibility in setting bud getary priorities and eliminate the possibility of later dis agreement between donor and institution regarding the use of a gift21 Explicitly exempting restricted gifts from the calculation of tax arbitrage under an expanded defini tion would increase the attractiveness of restrictions to schools but encouraging such restrictions could reduce the social benefit of charitable giving for instance individual donors may impose restrictions on gifts that diminish the public benefits that a school provides22 To reflect schools legitimate need for investment assets CBO calculated measures of tax arbitrage that would allow some investment assets to be set aside in a reserve that by definition could not earn returns from tax arbi trage In calculating those measures the value of invest ment assets considered in the arbitrage calculation was reduced by an amount equaling one year s operating expenses the median amount of investment assets held by schools issuing debt in 2003 CEO did not separately adjust the arbitrage estimates for investment assets subject to restricted uses reliable data on the type and strength of 20 Although no data exist that detail the exact purposes of accumu lated restricted endowment assets the Council on Aid to Educa tion collects information on the restricted purposes of current gifts to school endowments In 2008 the most common restriction for endowment gifts was for student financial aid accounting for 34 percent of restricted giving Athletics and academic depart ments each accounted for about 20 percent of restricted gifts to endowments 21 Disagreements about such restrictions have increasingly resulted in costly legal battles between schools and donors See John Hechinger New Unrest on Campus as Donors Rebel 1 Street journal April 23 2009 In the most prominent example of such a dispute a donor s heirs filed a lawsuit against Princeton University s Woodrow Wilson School of Public and International Affairs alleging that the school was not using a large donation dating to 1961 in accordance with the donor s intent Princeton paid a settlement of 100 million to end the ligation which had stretched on for six years 22 See Burton A Weisbrod Jeffrey P Ballou and Evelyn D Asch Mission and Money Understanding toe University New York Cambridge University Press September 2008 p 121 12 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES restrictions that apply to existing endowments are not available23 Measuring the Volume of Arbitrage Bonds Under a Broader De nition of the Term CBO used two main measures to determine how much of the value of outstanding bonds issued by colleges and universities and their new tax exempt borrowing could be considered to earn returns from tax arbitrage under a broader definition of the term Each approach compared an institution s investment assets with a measure of its tax exempt bond liability With the historical measure CBO considered only the outstanding stock in 2003 of previously issued bonds According to that backward looking measure if the value of a college s or university s investment assets exceeded the value of the institution s outstanding stock of tax exempt bonds all the bonds were classified as earning returns from tax arbitrage If the value of the stock of tax exempt bonds exceeded the value of the investment assets the volume of bonds equal to the value of the investment assets was classified as earning arbitrage profit The second more forward looking measure of tax arbi trage considered the position of colleges and universities in the years immediately following an expansion of the definition of arbitrage That first year measure com pared an institution s new issues of tax exempt bonds in 2003 with its stock of investment assets that year Under the first year measure if the value of investment assets exceeded the value of the new bond issues all of the bonds were classified as earning arbitrage profit If the value of the new issues of bonds exceeded the value of the institution s investment assets the amount of the new issues that was equal in value to the investment assets was considered to be earning profit from tax arbitrage In the first few years after the implementation of such a policy a relatively large share of new issues would probably be considered arbitrage bonds because the amount of invest ment assets newly available for yield restriction would be large compared with new issues in any single year Analogous to an estimate of tax expenditures the first year measure is not meant to capture all of the ways in which issuers might respond to a change in the definition 23 Of the 251 schools that borrowed in 2003 238 provided informa tion on temporary and permanent restrictions on their total net assets About 55 percent of all net assets were under permanent restriction and 24 percent were under temporary restriction No information is available on the nature of the restrictions of tax arbitrage It does however capture the immediate effect of broadening that definition24 Estimated Volume of Arbitrage Bonds Under a Broader De nition of the Term A substantial portion of the tax exempt debt issued by colleges and universities is outstanding at the same time those institutions hold higher yielding investment assets Such debt would earn profit from tax arbitrage under an expanded definition of the term that considered all investment assets not just those directly related to the tax exempt debt CBO s analysis indicates the following I If no reserve was allowed close to 100 percent of the outstanding tax exempt debt would be classified as earning full or partial returns from arbitrage under the broader definition If schools were allowed to exempt investment assets equal to one year s operating expenses as a reserve 72 percent of the currently out standing debt would be earning full or partial returns from arbitrage I Considering only new issues the share of debt with full or partial arbitrage profit would be about 99 per cent if no assets were set aside in a reserve and about 75 percent if a reserve equal to one year s expenses was allowed 24 CBO also calculated a related measure of arbitrage bonds not reported in the tables based on both outstanding bonds and new issues If schools continued to use tax exempt debt to finance cap ital projects and accumulate investment assets as they have in the past their new bond issues would tend to exhaust the assets sub ject to yield restriction lowering the share of new issues that would be earning arbitrage profits in later years If that was the case the current balance between the stock of investment assets and the outstanding stock of tax exempt bonds would remain the relationship between the two Under that assumption another way to estimate the amount of 2003 issues that would be consid ered arbitrage bonds would be to measure the new bond issues against the investment assets that remained after the outstanding stock of previously issued bonds was applied to those assets By that measure if the value of a new bond issue was less than the difference between investment assets and the outstanding stock of tax exempt debt the issue was classified as earning tax arbitrage If a new issue was greater than the residual investment assets the amount of the issue equal to residual investment assets was classi fied as earning returns from arbitrage By this definition the pro portion of debt that would be earning tax arbitrage was lower than that under the first year measure by about 10 percentage points Table 3 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES TaxExempt Debt Classi ed as Earning Pro ts from Arbitrage in 2003 Under a Broader De nition of the Term Percent No Assets Set Aside Assets Set Aside as a Reserve as a Reserve Historical FirstYear Historical FirstYear Measu rea Measu reb Measu rea Measu reb Debt Classified as Earning Arbitrage Profits Under a Broader Definition 100 99 72 75 Full Arbitrage 90 84 60 56 Partial ArbitrageC 10 14 12 19 Debt Not Classified as Earning Arbitrage Profits Under a Broader Definition l 28 25 Memorandum Total TaxExempt Debt Millions of dollars 22199 5703 22199 5703 Source Congressional Budget Office based on information returns Forms 8038 and 990 filed with the Internal Revenue Service by institutions that borrowed in 2003 Note less than 01 percent a The total volume of taxexempt bonds issued as a percentage of net investment assets adjusted for misreporting b Estimated volume of taxexempt bonds in the first year under a broader definition of tax arbitrage approximated by the volume in 2003 as a percentage of net investment assets adjusted for misreporting c Issues held contemporaneously with investment assets lower in value than the total issue Most of the debt that would be classi ed as earning returns from arbitrage is held by the schools that have the largest stocks of investment assets Compared with CBO s previous estimates of the volume of arbitrage bonds held by nonpro t hospitals institutions of higher learning hold considerably more debt that would be classi ed as earning returns from arbitrage under a broader de nition Estimated Amount of Arbitrage Debt The historical measure applies the broader de nition of arbitrage to already outstanding tax exempt debt Although the measure does not directly address the effect of a policy change because such a change would affect only future issues it provides a useful starting point for comparing the outstanding stock of tax exempt debt with investment assets Using data corrected both for misclassi ed tax exempt debt and for underestimates of investment assets owned by institutions that hold assets in foundations CBO found that in 2003 close to 100 percent of the 22 billion in previously issued tax exempt debt would be classi ed as earning returns from arbitrage under the broader de nition see Table 35 In other words almost all of the outstanding debt was issued by schools that also held higher yielding investment assets Furthermore the majority of the debt that would be classi ed as arbitrage debt was fully arbitraged that is almost all of it was issued by schools that held invest ment assets greater in value than their outstanding stock of bonds In the rst years after an expansion of the de nition of arbitrage a high proportion of new issues would be subject to yield restriction because accumulated invest ment assets are large relative to any single year s issues In 2003 about 99 percent of new bond issues would have been considered to be earning pro t from tax arbitrage under a broader de nition 25 The appendix presents alternative calculations of bond holdings that would earn pro ts from tax arbitrage under a broader de ni tion using no adjustments for misreporting tax exempt bond liability or for assets held by supporting organizations 14 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES Table 4 Colleges and Universities Conducting Tax Arbitrage in 2003 Under a Broader De nition of the Term Percent No Assets Set Aside as a Reserve Assets Set Aside as a Reserve Historical FirstYear Historical FirstYear Measurea Measu reb Measurea Measu reb Institutions Conducting Arbitrage Under a Broader Definition 89 98 44 52 Full arbitrage 66 82 31 37 Partial arbitrageC 23 16 14 15 Institutions Not Conducting Arbitrage Under a Broader Definition 11 2 56 48 Memorandum Total Institutions 251 251 251 251 Source Congressional Budget Office based on information returns Forms 8038 and 990 filed with the Internal Revenue Service by institutions that borrowed in 2003 a Total volume of taxexempt bonds issued as a percentage of net investment assets adjusted for misreporting b Estimated volume of taxexempt bonds in the first year under a broader definition of tax arbitrage approximated by the volume in 2003 as a percentage of net investment assets adjusted for misreporting c Issues held contemporaneously with investment assets lower in value than the total issue The amount of debt that earns returns from arbitrage falls when some assets are set aside as an exempt reserve Using the historical measure which compares the outstanding stock of tax exempt bonds to investment assets not set aside for a reserve the amount of debt earning arbitrage pro ts in 2003 fell to about 16 billion or about 72 percent of the stock of outstanding tax exempt debt For issues in 2003 about 75 percent of new tax exempt bond issues would be classi ed as earning returns from tax arbitrage under a broader de nition that allowed an 2 exempt I39CSCI39VC equal to OHC year s CXpCHSCS 6 26 By that measure the total amount of debt that would be classi ed as earning returns from arbitrage under a broader de nition is about the same if the exempt reserve is calculated as 70 percent of investment assets rather than as one year of operating expenses However the distribution of arbitrage debt among the schools would differ De ning the exempt reserve on the basis of assets rather than expenses lowers the amount of arbitrage debt for schools with assets that are large in relation to expenses and increases it for schools with assets that are small in relation to expenses In other words de ning an exempt reserve on the basis of assets favors schools that have better access to taxable borrowing The Distribution of Arbitrage Debt The percentage of institutions engaging in the practice of tax arbitrage is generally lower than the percentage of debt that generates arbitrage pro ts because the 50 institutions with the most investment assets account for a large share of that debt If the expanded de nition of arbitrage was applied to the already outstanding stock of tax exempt debt about 90 percent of institutions would be characterized as pro ting from the practice of tax arbitrage see Table 4 According to the rst year mea sure nearly all of the institutions issuing new debt in 2003 would be classi ed as earning arbitrage pro ts if a broader de nition had been in effect that year Even if the broader de nition of arbitrage allowed col leges and universities to set aside substantial investment assets in a reserve the majority of tax exempt debt held by those institutions would be classi ed as earning returns from tax arbitrage regardless of the measure used Because the distribution of arbitrage earnings is not uni form however if a reserve was allowed only about half of the tax exempt colleges and universities that borrowed in 2003 would be Viewed as conducting arbitrage at all under either measure Under the historical measure the number of institutions conducting arbitrage would fall Figure 3 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES The Distribution of TaxExempt Debt Held by 251 Colleges and Universities in 2003 That Would Be Classi ed as Earning Profits from Arbitrage Under a Broader Definition of the Term Percent 80 7o I Bonds Earning Arbitrage Profits Using CBO39s FirstYear Measure With No Assets in a Reserve 60 Bonds Earning Arbitrage Profits Using CBO39s FirstYear Measure With Assets in a Reserve 50 40 30 20 10 0 1 to 10 11 to 50 51 to 100 101 to 150 151 to 200 201 to 251 Institutions Ranked According to Reported Investment Assets Source Congressional Budget Office Notes CBO used two main measures to determine how much of the outstanding stock of bonds held by colleges and universities and their new taxexempt borrowing could be classified as earning profits from tax arbitrage under a broader definition of the term With the historical measure CBO considered only the historical outstanding stock of previously issued bonds The second more forward Iooking measure considered the position of colleges and universities in the years immediately following an expansion of the definition of arbitrage Allowing some assets to escape the broader definition of arbitrage lowers the share of debt considered to be earning arbitrage profits by schools with lower amounts of investment assets from 223 to 111 or to about 44 percent of all 2003 bor rowers if some assets were set aside in an exempt reserve Using the rst year measure based on the volume of new issues 52 percent of institutions would have been classi ed as engaging in arbitrage with new issues in 2003 after accounting for an exempt reserve Broadening the de nition of tax arbitrage would affect institutions with large holdings of investment assets more than institutions with fewer holdings The majority of bonds that would be classified as earning arbitrage profit using the rst year measure were issued by the institu tions with the largest investment assets The 50 borrowers with the largest investment asset balances which com prised about 20 percent of all schools that issued new tax exempt debt in 2003 accounted for just over 60 percent of the bonds earning tax arbitrage see Figure 3 If assets equal in value to one year s operating expenses were set aside about 80 percent of the bonds earning returns from arbitrage would be on behalf of the 50 borrowers with the largest holdings of investment assets Comparison With Arbitrage Debt for Nonpro t Hospitals In a previous study CBO used similar methods to esti mate the percentage of nonprofit hospitals and their debt that would be classified as earning returns from tax arbi trage under an expanded de nition By every measure a much larger percentage of schools than nonpro t hospi tals would be conducting arbitrage under an expanded de nition and a larger share of debt issued by colleges and universities than debt issued by nonpro t hospitals would be classi ed as earning arbitrage pro ts Using data adjusted for misreporting CBO found that about 60 percent of the outstanding stock of tax exempt 27 Congressional Budget Of ce Nonpro t Hospitals and Tax Arbitrage 15 16 TAX ARBITRAGE BY COLLEGES AND UNIVERSITIES bonds issued by nonpro t hospitals in 2002 had been issued by hospitals that contemporaneously held higher yielding investment assets Using the first year measure 64 percent of bonds issued by nonprofit hospitals in that year would be classified as earning arbitrage profits under a broader definition For colleges and universities those figures were close to 100 percent Because different criteria are used to rate hospital bonds and those issued on behalf of colleges and universities CBO used a different method in its previous study to cal culate the amount of assets that might reasonably be set aside as a reserve28 According to those estimates if an exempt reserve had been allowed under an expanded def inition of tax arbitrage in 2002 about 33 percent of the outstanding stock of bonds issued by nonprofit hospitals would be earning arbitrage profits in that year Under the first year measure and assuming that an exempt reserve for hospitals would be allowed CBO determined that 32 percent of new issues of tax exempt bonds would be classified as earning arbitrage profits under a broader def inition The corresponding figures for colleges and uni versities were over 70 percent Institutional Response to an Expanded Definition of Tax Arbitrage Schools could adjust to a broadened definition of arbi trage in several different ways They could issue fewer bonds and reduce their capital spending They could sell or reduce their stock of investment assets in order to finance capital spending rather than issuing tax exempt debt and rebating the yield on investment assets to the federal government They could also replace borrowing that would result in yield restriction with taxable debt In all cases the net cost of the tax preference to the federal government would be reduced Another possibility is that decreasing the attractiveness of tax exempt private activity bonds might encourage insti tutions of higher learning to pursue tax exempt financing through other channels For example when the Tax Reform Act of 1986 limited the use of private activity bonds for the financing of sports stadiums agreements between local governments and sports teams led to the issuance of bonds specifically designed for stadium financing but legally considered general revenue bonds because they were backed by revenue from general sources In those cases the requirement that debt service 28 The reserve was set equal to 100 days of operating expenses the median hospital in the sample had 117 days of cash on hand be paid from revenues not generated by the stadium essentially ensured that local governments would offer teams very favorable lease terms and that the tax burden of such facilities would be shared by all local taxpayers rather than the users of the facility who most benefited from it29 Local governments could allow both public and private universities to circumvent expanded tax arbitrage rules for private activity bonds in a similar fashion Although colleges and universities do not enjoy the monopoly power of major sports teams many institutions of higher learning are large landowners and employers at the local level giving them substantial negotiating power with local governments Local governments could also desig nate public schools as separate units of government for the purpose of issuing bonds allowing those schools to issue general revenue bonds on their own Although care ful regulation could in theory circumvent that problem it is likely that at least some issuance of tax exempt private activity bonds would simply shift to the issuance of tax exempt revenue bonds limiting the revenue gain to the federal government Broadening the rule would reduce the tax preference for schools with large asset portfolios to a greater degree than it would for schools with fewer resources Because the new rule would be more likely to apply to the few schools with very large portfolios those schools would effectively face an increase in interest costs relative to schools with smaller portfolios However those schools with signifi cant investment assets already tend to have investment grade credit ratings suggesting that limiting their ability to issue tax exempt debt would probably not prevent them from accessing financial markets Whether divert ing the tax exempt bond subsidy away from schools with larger endowments would be a more efficient use of scarce resources would depend on the marginal social benefit of subsidies to such schools If the marginal social benefit of the subsidy decreased as endowments rose a reduction in the subsidy to schools with large endow ments could improve the allocation of the nation s resources If the marginal social benefit of the subsidy increased as endowments rose a reduction in the subsidy to schools with large endowments would worsen resource allocation 29 Dennis Zimmerman Subsidizing Stadiums Who Benefits Who Pays in Roger G Noll and Andrew S Zimbalist eds Sports jobs and Taxes toe Economic Impact ofSports Y ams and Stadiums Washington DC Brookings Institution Press 1997 pp 1 19 145 Appendix Alternative Calculations of Tax Arbitrage as Practiced by Colleges and Universities In the main text of this report the Congressional Budget Of ce CBO calculated the extent of tax arbi trage that colleges and universities would be practicing under an expanded de nition of the term To do so CBO adjusted the outstanding stock of tax exempt bonds held by those institutions to reflect the fact that they often misreported such debt as mortgages and expanded their measured investment assets to include assets held by related organizations In this appendix CBO presents estimates based on the same data taken directly from information returns specifically Form 990 returns led with the Internal Revenue Service IRS but without making such adjustments According to unadjusted data from Form 990 returns which nonprofit entities are required to submit to the IRS on an annual basis the percentage of previously issued outstanding debt that would be classi ed as earn ing returns from arbitrage under a broader de nition would be similar to the percentage produced using adjusted data In determining those percentages CBO used the historical stock of previously issued bonds which it terms the historical measure However 148 institutions slightly less than 60 percent of the 251 schools that borrowed in 2003 had investment assets with a value that exceeded the reported stock of tax exempt liabilities a considerably smaller percentage than was the case when adjusted data were used Most of the additional stock of debt added by the adjustment was for institutions that reported no tax exempt debt at all on Form 990 returns but the additional stock of bonds attributable to the adjustment would be small relative to the total stock of bonds Estimates that were produced using what CBO terms the rst year measure which considers the position of colleges and universities in the years immediately following an expansion of the de nition of arbitrage do not include information on the outstanding stock of tax exempt bonds so they are affected only by the adjust ment to investment assets Using that measure and unadjusted data CBO determined that the share of debt earning returns from arbitrage would be slightly lower by about 1 percentage point than when adjusted data were used Using that measure and unadjusted data the num ber of institutions conducting arbitrage would be slightly lower as well by about 2 percentage points Again both the adjusted and unadjusted data show that a majority of the new issues of tax exempt bonds for institutions of higher learning would be classi ed as earning arbitrage pro ts after an expansion of the de nition of tax arbitrage CONGRESS OF THE UNITED STATES CONGRESSIONAL BUDGET OFFICE WASHINGTON DC 20515 INSIDE MAIL
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'