The IRS Business Master Files include a field (RULEDATE) indicating when registered nonprofit organizations obtained formal recognition of their tax exempt status by the IRS. (In other words, when the IRS approved their applications for exempt status.) NCCS typically uses this as a proxy for when an organization was created. However, one should understand its origins and flaws before determining how best to use it in one's research.
Limitations are of two types:
- Nonprofit corporations -- the majority of organizations -- must incorporate before they register with the IRS. This could occur at more or less the same time as they file with the IRS, but may also occur a year or more earlier. (Incorporation is handled by state governments.) Moreover, some organizations begin informally without any formal legal structure. Thus, depending on one's definition of "founding," the ruling date may or may not be adequate as a proxy.
- Prior to the 1960s, IRS nonprofit information was maintained in paper form only and recording of ruling dates appears to be spotty. This is discussed at greater length below.
IRS files were computerized in the 1960s. Prior to that time, collection of ruling date information appears limited. Only 138 501(c)(3) organizations out of more than 800,000 registered in 2002 had ruling dates prior to 1920.
A review of major private universities shows that many have 20th Century ruling dates that clearly do NOT reflect their date of establishment. An additional 8,884 (1%) had no ruling dates entered. (These appear as '0000' on the file. An analysis of the 2004 Core Public Charity File shows that approximately the same percentage (0.9%) of Form 990 filers have no ruling dates, too.) However, universities are NOT typical of the sector as a whole, which includes a much larger percentage of newer organizations.
To put this in perspective, 41% of organizations in the higher education NTEE categories (B4 and B5) had ruling dates from 1970 or later. (This surprisingly large number includes a range of entities, such as schools or special programs within existing institutions as well as some misclassified entities.) In comparison, more than twice that percentage -- approximately 85% of all 990-filing public charities -- had ruling dates of 1970 or later, and 92% of these are for 1960 or later. (Analysis of the 2004 NCCS Core Public Charity File)
For additional discussion of the limitations of this variable, see the mid-2000s article in Nonprofit and Voluntary Sector Quarterly by Kirsten Gronbjerg, et al., comparing ruling dates and incorporation dates for Indiana nonprofit organizations for further discussion.
TWO APPROACHES TO USING THE RULING DATE IN REGRESSION ANALYSIS
There are at least two options for dealing with the limitations of this variable:
1) Recode the ruling dates into age ranges -- pre-1970, 1970-80, 80-90, 90-00 (or possibly 5-year increments in later years). This makes the impact of age harder to interpret but seems a safe approach.
2) "Bottom-code" age at, say, 1970. Thus, any entry (including missing entries) would be coded as "1970."