Using NccsDataResources For NonProfit Research

Using NCCS Data as a Tool for Nonprofit Research September 29, 2006 Tom Pollak Associate Director National Center for Charitable Statistics at the Urban Institute http://nccs.urban.org The National Center for Charitable Statistics at the Urban Institute • Repository of IRS data on nonprofit organizations; regular research reports, Almanac on dimensions of sector • NCCS created 1981; later, part of Independent Sector’s research division • Moved to Urban Institute in 1996 – Nonpartisan economic & social policy research organization • NCCS is part of Urban’s Center on Nonprofits and Philanthropy headed by Elizabeth Boris NCCS – The Past 10 Years • Small team of staff & consultants • Leverage through technology & collaborations • Building data infrastructure – Collecting & standardizing files – Putting data online – Providing guides and resources to make data accessible – The Foundation Center NCCS – The Past 10 Years (2) • Data quality – Electronic Filing of 990s – Annual IRS-NASCO Form 990 Meeting – Overhead Cost Study & other research • Building new data sources – GuideStar & IRS collaborations – State nonprofit associations, umbrella orgs – The Foundation Center Nonprofit Organizations Concept Map Registered with IRS but not filing annually Registered 501(c)(3) Private Foundations All priv. foundations must register & file annually Registered Other exempt organizations 501(c)(4), 501(c)(5), etc. 501(c)(3) Public Charities Registered & Filing Annually Registered & Filing Annually Registered AND filing annual returns Informal & unregistered organizations & community groups (numbers unknown) Dimension #1: IRS Registration • Are the organizations registered with the IRS? – A neighborhood association, recreation club, or community theater with gross receipts less than $5,000 per year is a nonprofit but is not required to register with the IRS. • Religious congregations are not required to register, but many do so anyway. Dimension #2: Annual Form 990 Filing • Is organization required to file an annual IRS Form 990? (Gross receipts are more than $25,000 & org. is not a congregation.*) • All 501(c)(3) private foundations must file annual IRS Form 990-PF forms. * Congregations, religious denominations, & closely related organizations are exempt from filing. Dimension #3: 501(c)(3)s and Other Exempt Orgs. • Are donations to the organization tax-deductible? – Public Charities. If donations are fully deductible, then likely to be a “501(c)(3) public charity.” These include most nonprofit arts groups, universities, hospitals, & human service organizations. – Private Foundations. If donations are partially deductible, then likely to be a “501(c)(3) private foundation.” These include the Gates Foundation, Ford Foundation, Carnegie Corporation, & 60,000 others. • Otherwise, it could be a trade association, labor union, Elks Club, HMO, or a wide range of other “exempt organizations.” Dimension #4: Formal - Informal • There is no agreement on the outer boundaries of nonprofit “organization” • The Formal: – Everyone agrees that a private college that files an IRS Form 990 is a nonprofit. The Less Formal Are these “nonprofit organizations”? • A neighborhood association that only meets every couple of years and is NOT incorporated? • A group of people that get together for a 6-month fight against the building of a Wal-Mart? • A group of people who get together once a month for football or to play jazz? • What if they call themselves the “Podunk City Jazz and Football Club”? Have a club president? Pay dues? Advertise for new members? Are incorporated? Datasets Available • IRS Business Master Files • NCCS Core Files, based on Form 990 • GuideStar-NCCS Nonprofit Org Database, based on Form 990 • IRS Statistics of Income Samples • Other Files... Good News, Bad News • No perfect file. All have limitations • If clear about research questions and goals, can usually make good use of our data IRS Business Master File • All organizations registered as tax exempt with IRS • 1.4 million (May ’06) • Files available for 1989, 1995-2006 • Basic descriptive information – Mostly from Form 1023 Application for Exemption – Gross receipts, assets, tax period from most recent Form 990 BMF Limitations • Missing years in early 1990s • 21% of orgs couldn’t be located in a 1990s survey • Only the most basic descriptive information • Addresses may be out-of-date • Some errors on finances • Only partial (and probably not representative) coverage of religious organizations NCCS Core Files • Annual files of organizations filing IRS Forms 990 • Standard methodology: 1989-2004 • Public charities, private foundations, and non-501(c)(3) organizations • 303,000 in 2004 • 100+ variables • Major financial variables, most types of rev., key expense detail NCCS Core Trend Files • “Long and narrow” compilation of small subset of key financial variables from 1989 through latest Core file • Available for public charities or private foundations • Update for 2004 should be available soon. • We can also create custom “wide” datasets with imputed values for missing years Core File Limitations • 100 out of 400 or so possible variables • No government grants! • Most revenue vars, but weak detail in almost all other sections • Missing returns: 1990, 1992… • No “zero filers” (those reporting <$25k in gross receipts) – non-issue for most purposes GuideStar-NCCS Nonprofit Organizations Database: “Digitized Data” • Key punched independently by GuideStar from images scanned by IRS and purchased by NCCS • Finances checked and edited by NCCS • NCCS adds classifications, and other fields for researchers • All 501(c)(3) public charities filing 990s, EZs • FY 1998 – 2003 “Digitized Data” (2) • • • • • 400+ variables in multiple files Part 3 program & org. descriptions Part 5 and Sch. A, Part 1 compensation Gov’t grants & full revenue detail Full functional expense & balance sheet detail • “Naughty bits”: Other information and Sch. A “Digitized Data” Limitations • Only public charities • Approximately 1% of EINs don’t match IRS’s official list – no easy way to merge with or supplement Core Files • Data quality issues in Part 2 due to poor reporting by charities • Missing organizations in 1998 due to IRS scanning problems • Challenge working with large, complex database IRS Statistics of Income (SOI) Samples • Annual files of organizations filing IRS Forms 990 -- 1982-2003 • Public charities, private foundations, and non-501(c)(3) organizations • 16,000 organizations in 2002 public charity sample (990 filers plus & 500 EZ filers) • 400+ financial variables for 990 filers SOI Samples (2) • All orgs with $30+ mil. in total assets • Sample of others, and very few EZ returns • Skilled editors adjust rev., exp. & bal. sheet line items • Ideal for national studies of finances of large organizations SOI Limitations • Field names (“P010”) difficult for researchers (although easier for “array processing” in SAS) • Editing/changing of items is both a strength and a weakness • No access to confidential 990-T information on unrelated business income without “special sworn status” • Little/no value for geographic or state analysis Other Files • Unified Database of Arts Organizations – combines IRS, state arts agency mailing lists, and other sources • Itemized household giving by zipcode, 1997 • “501(h) electors” – Organizations that report their advocacy expenses under special IRS rules • 1992 program descriptions from IRS 990 Parts III & VIII NCCS Data Web NCCS Resources for Researchers http://nccs.urban.org • • • • • • Researcher’s Guide to NCCS Data Database file descriptions KnowledgeBase, FAQ+ include technical notes, “snippets” of information, and more Table Wizards, fact sheets and state profiles The DataWeb Classification Tools Getting Access to NCCS Data • The problem with intellectual property: Marginal costs for distribution is zero but cost to create is large • NCCS mission is to make data available for research purposes • Until several years, we made all data available for free to researchers, but can no longer afford this “business model” NCCS Data: Our Costs • Tranforming raw data into researcherfriendly databases • Cleaning the financial data • Purchase of rights to GuideStar data • Ongoing classification • Processing data requests • Technical support to users & the IRS, Creating & maintaining the DataWeb • Outreach & data quality work • And more… NCCS Data: Your Costs • If seeking funding for research using our data, PLEASE talk to us first about including real cost of data in your funding proposals • DataWeb access but no data extracts: • Free for students and scholars for research purposes • Cost for Digitized Data – nominal • For unfunded projects, typical cost in the $100500 range depending on number of databases, with discounts based on need (“the impoverished student” loophole) Classification • NTEE organization-level classification – National Taxonomy of Exempt Entities – Purpose, type and major function of orgs. – Translates to federal government’s more general NAICS codes • Nonprofit Program Classification (NPC) – Program level activity codes – 1 mil.+ programs – NCCS implemented using auto-coding systems Classification Quality • NTEEs – 90% at major group level – All records given a NTEE confidence code – 800,000 records with firm codes, another million with “draft” auto-codes • IRS Activity Codes – NCCS discourages use – Old IRS self-coding system – Useful information – 3 per org. assigned: 1st is best NTEE Classification • • National Taxonomy of Exempt Entities (NTEE) created in early 1980s Hierarchical: – – – 10 major categories: From arts, environment, education, health care, human services, & lots of “others” Broken into major groups (letters of the alphabet) 600+ “decile” and “centile” codes Major group “P”: Human Services (general) “P6”: Emergency Assistance “P62”: Victims’ Services • Example: – – – NTEE Classification • • • • • • National Taxonomy of Exempt Entities (NTEE) created in early 1980s Used by NCCS, Independent Sector, the Foundation Center Starting in 1998 IRS “determination specialists” began assigning codes to new organizations – quality an issue IRS SOI classifications started independently around the same time for SOI sample NTEE-XP used by Foundation Center “Condensed” NTEE-CC (only 600 codes and no “common codes”) created for IRS by NCCS NTEEs A - Arts, Culture, and Humanities B - Education C - Environmental Quality, Protection, and Beautification D Animal-Related E - Health F- Mental Health, Crisis Intervention G- Diseases, Disorders, Medical Disciplines H - Medical Research I - Crime, Legal Related J - Employment, Job Related K - Food, Agriculture, and Nutrition L - Housing, Shelter M- Public Safety N - Recreation, Sports, Leisure, Athletics O - Youth Development P - Human Services - Multipurpose and Other Q - International, Foreign Affairs, and National Security R - Civil Rights, Social Action, Advocacy S - Community Improvement, Capacity Building T - Philanthropy, Voluntarism, and Grantmaking Foundations U - Science and Technology Research Institutes, Services V - Social Science Research Institutes, Services WPublic, Society Benefit Multipurpose and Other X - Religion Related, Spiritual Development Y - Mutual/Membership Benefit Organizations, Other Z - Unknown/Unclassified EON Online Classification Tool Getting Started with the DataWeb • • • Become a registered user at http://nccs.urban.org or email nccs@ui.urban.org One-Minute Tours on the Data Web Guide to Using NCCS Data “The Worst” NCCS & IRS E-Filing • July 1999 – Began e-filing project (IRS schedules e-filing of 990s for 2007) • Feb. 2002 – Live in Pennsylvania with Desktop990 (990 and state forms) • Oct. 2002 – Colorado added • Feb. 2004 – IRS goes live with e-filing 990s; NCCS unveils 990-EZ Online • Feb. 2005 – NCCS unveils 990 Online. To date, nearly 9,000 returns created. • Jan. 2006 – IRS requires largest orgs to e-file, implements Fed-State Retrieval System, 7 certified 990 e-filing providers • March 2006 – New infrastructure supports multiple states, state-only filings, & morec 990 Online: The TurboTax for Nonprofits! • Web-based system for preparing Form 990 and e-filing with IRS • Immediately improves quality of reporting • More than 9,000 returns created • Free • Easy to use • Drop-down menus, calendars for selecting items & dates • Program takes care of all arithmetic • Verifies that all required items are included • Context-sensitive links to IRS help and expert tips 990 Online • • • • • Generates an Adobe PDF for easy printing, saving Electronically file with the IRS Adding state forms now Secure Import common schedules from Excel • Data in usable format • Saves data from year-to-year • No duplication of entries: Questions are asked only once http://efile.form990.org 990s, EZs, Extensions… Import Information from Last Year’s Return 50+ Attachments, No Math Create a 990 That Your Organization Will Be Proud Of The Future of E-Filing • Integration of federal and state e-filing systems • Dramatic improvements in quality of information that nonprofits put on forms • Dramatic improvements in ability of regulators to ensure quality of data by checking for missing items, checking ratios, etc. • Decline in costs of communication from regulators if can be done electronically • Save tens of millions in key punching costs • Open door to better use of 990s for outcome measurement, use by umbrella associations, and researchers New Projects: Helping New Organizations Register with the IRS • NCCS working on development of webbased Form 1023 Cyber Assistant – application for tax-exempt status – for IRS • Provides assistance for preparer, like 990 Online • Goal is better quality applications; also usable data Using 990 Data for Financial Analysis: Themes • “All 990 Data is Not the Same”: Some 990 items are highly reliable and valid; others are not • No single “nonprofit sector”: Almost always need to distinguish large & small • Sector vs. panel analysis • Fiscal vs. “Core”/”circa” years No Single Nonprofit Sector: Distinguishing Large and Small Orgs in Analyses • Financial profiles are very different • If aggregate large and small,large will overwhelm and distort small orgs’ finances • Combine medians,ratios, and panels instead of aggregating dollars Good Parts, Bad Parts • • • • • • Part 1 Revenues: B+ Part 2 Functional Expenses: C Part 4 Balance Sheet: AParts 4-A & B: APart 5 Compensation: B Part 7 Income-Producing Activities: C Part I: Revenue • Gross receipts vs. total revenue • Government grants or contracts? • Audited financials’ “grants and contracts” • Statutory need to distinguish & redefine government grants • Read the instructions! • Private contributions • Dues • Special events Part I: Net Income • “Excess (or deficit) for the year” – line 18 • Beware of this number: Unstable from year to year • The effect of multi-year grants, inkind contributions, etc. Part II: Functional Expenses • 990,not EZ • What’s fundraising? • 37% of orgs reporting more than $50k in direct private contributions reported ZERO fundraising costs! • Urban Inst./Indiana Univ. Center on Philanthropy Overhead Cost Study • http://www.coststudy.org • Guidelines clear but POOR reporting, understanding Part II: “Other Expenses” • More than 30% of "other expenses" could be allocated to one of the existing lines on the Form 990. • More than 1/3 of the amount reported by the sample in "other expenses" were for consulting services. • Other major categories included occupancy costs (already on the 990), advertising, and insurance. Part IV: Balance Sheet • Best section on 990 • Closely corresponds to audited financials • Weaknesses: No distinction between current and long-term assets and liabilities • No “quasi-endowments” or boarddesignated funds distinguishable • Use investment assets as proxy? • Liquid assets: all but real estate and equipment? Part IV-A & B: Reconciliation with Audited Financials • • • • • Can be very useful if small sample On latest SOI file? Not in Digitized Data Donated professional services Unrealized gains and losses from investments Part VII: Income Producing Activities • Gross under-reporting of Medicare and Medicaid revenue. Why? • Nobody wants to pay Unrelated Business Income Taxes • Only available on SOI files “Core/Circa” vs. Fiscal Year • Fiscal year: Only returns with ending period in given fiscal year: 12/31/2005 and 6/30/2005 = FY 2005 • Core/Circa Year: Latest return received within past 2 years • SOI Year: (Almost) only returns STARTING in given calendar year. • Choice depends on research question, availability of data Geographic Analysis • Street Addresses – 23% are PO Boxes in BMF – 27% in Core files – All but several hundred have 10-digit zipcodes • County FIPS codes, using 5 digit zipcodes • MSAs derived from FIPS Challenges • Headquarters vs. service delivery area • National offices, local branches • Community good or ill? – Group homes vs. services for residents • What’s missing? – Depends on the study – Missing food pantries, small arts orgs – Religious service organizations – maybe – Government services Using NCCS Data for Surveys • Mail surveys – 21% of BMF orgs. cannot be located (nonfilers) – GuideStar “digitized data” vs. Core Files – Use GuideStar data for contact names • Phone surveys – Use GuideStar data for phone numbers & contact names – Challenges reaching large organization Exec. Directors Which Sampling Frame? • Start with population, and start excluding… • Include unstaffed organizations? (personnel costs as indicator) • Include trusts? • Include supporting organizations? • Include orgs with national or non-service focus? • Start with narrowly tailored group of operating organizations? Response Rates • 50% possible, but very difficult and expensive • Pre-Tests and Cognitive Interviews • Dillman multi-modal approach: repeated mail & phone contacts • Use of incentives? • Internet option is easy, but don’t depend on it • NCCS weakness: No email addresses! • But outside services are option to explore