525 Annual Reports to Shareholders: Historical Collections in Libraries Judith M. Nixon Judith M. Nixon is Professor of Library Science in the Humanities, Social Science and Education Library at Purdue University; e-mail: jnixon@purdue.edu. © Judith M. Nixon The purpose of this article is to describe the scope and depth of the historic corporate annual report collections in twelve academic/research libraries in North America. For many decades, a few major academic business libraries have been collecting and preserving corporate Annual Reports (ARS), the reports sent to shareholders documenting the financial status and future plans of the company. Today these historic collections provide more than a record of the companies’ finances; they provide a glimpse into the social and cultural thoughts from the past and the corporate stories of individual companies. Digitization of these reports has only just begun; so, except for fewer than 900 companies (most of which have been on the Fortune 500 list), the companies included in these historic annual report collections are not available electronically. The combined collection of the twelve libraries includes reports from nearly 38,000 different companies. Stanford, Harvard, and Western Ontario have the largest collections. Har- vard and Columbia have collections with significant historic depth, while Purdue has a collection with important late–twentieth-century holdings. The overlap of the collections is much smaller than was anticipated. The two largest collections, Stanford with over 18,000 companies and Harvard with over 11,000 companies, have only 3,668 companies in common. The overlap of companies between the largest five collections is only thirty- three companies. This research identifies these collections as unique and therefore valuable to the study of specific company histories and the industrial development in North America. Librarians need to preserve these collections and work toward digitizing them. or nearly 180 years, compa- nies have been sending share- holders an annual report on the financial situation of the company. Ever since business colleges have existed, librarians have been collect- ing these annual reports to shareholders, commonly referred to as ARSs. Although heavily used by students and research- ers in university libraries in the past, the size, depth, and overlap of these collec- tions have never before been measured. The purpose of this research paper is to describe the scope and depth of the col- lections at twelve major research libraries that have collected printed annual reports and to measure the overlap of these col- lections. The twelve libraries included in this study are: 1. Angelo Bruno Business Library at the University of Alabama 2. Baker Library, Harvard University crl-56r1 526  College & Research Libraries  November 2010 3. C.B. “Bud” Johnston Library, Uni- versity of Western Ontario 4. Dewey Library for Management and Social Sciences, Massachu- setts Institute of Technology (MIT) 5. J. Hugh Jackson Library, Stanford University 6. Lippincott Library, University of Pennsylvania 7. Management & Economics Li- brary, Purdue University 8. Management Library, Cornell University 9. Science, Industry & Business Library of The New York Public Library (NYPL) 10. Thomas J. Long Business Library, University of California, Berkeley 11. Watson Library of Business & Economics, Columbia University 12. Yale University Libraries, Yale University Throughout the article, libraries will be referred to by the bold words in the preceding list. Literature Review Only two major research studies of the ARS reports in North American libraries exist: one from 1986 and the other from 2000. Both studies focus on the impor- tance of the reports to libraries, not the size or depth of the collections. Twenty- three years ago, Judith Bernstein con- ducted the first survey of 500 academic libraries and found that 75 percent of them collected printed ARSs, with most focusing on Fortune 500 companies. She recommended that major university li- braries maintain permanent archives but that most schools need only retain the reports for five years.1 In a similar follow- up survey of 121 Association of Research Libraries in 2000, Lisa O’Connor reported that libraries still found the convenience of the printed report valuable despite the fact that digital reports were available, with 55 percent still collecting printed reports.2 The major reasons for this drop in percentage were the availability of electronic financial information in com- mercial databases and the availability of ARSs on company Web sites on the World Wide Web. An important finding of these two re- search projects is that only a few of these libraries have archived their printed ARS reports while most retained them for less than five years. Bernstein’s survey report found that thirty-one libraries retain Fortune 500 company reports indefi- nitely, and O’Connor found that only ten libraries retain ARS reports indefinitely. (O’Connor surveyed fewer libraries but the set of large U.S. academic libraries was very similar in both surveys.) So between 1986 and 2000 the number of li- braries collecting ARSs had dropped, but the more important point for this article is that the number of libraries archiving the printed reports had dropped from thirty-one to ten. In the field of library science, besides these two research articles, there are only three other articles on ARSs, and none are investigations of the size and depth of the collections. Martha Jo Sani’s article, “Does the Corporate Annual Report have a Future?” written in 1991, includes a brief history on ARSs and outlines the useful- ness of the reports as a communication tool for companies.3 Steven J. Bell’s article, “Corporate Change: Impact on Corporate Documents Collections,” discusses the problems that company name changes cause for libraries in their efforts to orga- nize these reports.4 The most recent article authored by Cynthia L. Cronin-Kardon discusses the digitization of fifty annual reports at the University of Pennsylvania.5 So, despite the fact that many libraries have invested staff time and library space to ARS collections, research on these col- lections is sparse. Current Status and Usefulness of Archives of Historic Annual Report Collections in Electronic and Print Format Based on the research of Bernstein and O’Connor, between 1986 and 2000 at least twenty-one libraries have stopped Annual Reports to Shareholders  527 archived collections and in several cases withdrawn the reports. The major rea- sons given for discarding these reports is the perception that they are available in electronic format or that they are available in other libraries. However, this research study indicates that both these percep- tions are inaccurate. The uniqueness of the existing ARS collections is this article’s focus and will be discussed at length later in this article. As to the perception that these reports are available in electronic format, in real- ity a major percentage of these historic reports have not been digitized. Between the late 1990s and 2003, Lippincott Library at the University of Pennsylvania ex- perimented with scanning and digitizing ARS reports from fifty ARS companies. This project is described in an article by Cynthia L. Cronin-Kardon and Michael Halperin.6 Very recently Columbia Uni- versity announced a digitization project called “Historical Corporate Reports Online Collection,” which has the ARS reports from thirty-six New York City– based companies from the Columbia Business Library’s collection ranging from the 1850s through the early 1960s.7 Both of these projects include a very small percentage of the reports in either library. ProQuest Historical Annual Reports is the only large database that has digitized pre-1996 reports. This database preserves in electronic format the reports of the For- tune 500 companies dating back to 1844. It is of enormous importance to history scholars, especially those interested in the development of U.S. companies and industries; however, it is limited to those companies that were listed on the Fortune 500, about 800 companies. The major re- search libraries that have collected ARS reports have much deeper collections. It is true that much of the information in ARSs since the1980s that is used for financial analysis can be obtained more easily from databases that extract the data from the reports and repackage it in financial databases such as Mergent Online, Compustat, or Standard and Poor’s Net Advantage, so researchers in the fields of finance and accounting do not need to use the ARS as they did in the past. How- ever, like newspaper articles, ARS reports are primary research resources for historic information. Historians are not generally searching for financial data; they are look- ing for social, economic, and cultural in- formation. Since the reports were written in laymen’s terms, clearly identified the companies’ goals, and included pictures and charts, they have become a major source of information on the corporate stories of individual companies and the development of the U.S. industries. The fact that they were intended to be public relation tools from the companies only en- hances their usefulness to historians: One can use them to trace the impact of his- torical events on companies, industries, and the entire culture. Examples include the country’s response to World War II, including such social changes as women entering the workforce in huge numbers and how companies supported the war efforts. Another example is the changing preferences of the consumer and how companies marketed to the consumer. For example: the Pillsbury Doughboy, known as Poppin’ Fresh, is chronicled in their ARS report. Each report gives us a glimpse into the past, illustrating what each company was doing that year, and how it wanted the public to perceive it. A look at one re- port, Columbia Broadcasting Service Annual Report for 1950 will illustrate this. Long before color television was a consumer product, research was in progress. … (even though the Korean War had begun): “Your companies 10-year effort to bring color television to the public came to a climax when the Federal Communications Commis- sion adopted the CBS color system as standard for the industry.”8 The same 1950 Columbia Broadcasting Service Annual Report reveals that Steve Allen, Jack Benny, Burns and Allen, Gary 528  College & Research Libraries  November 2010 Moore, Ken Murray, and Frank Sinatra all made their debuts that year on television. Sports television was changing, too. Films of a Saturday game were developed on Saturday night, edited and scripted Sun- day morning, and shown early Sunday afternoon with Sports Director Red Barber pointing out key plays and developments. Using such information, historians can trace many trends and developments if they have access to strong collections of ARS reports. Compiling the List of ARS Reports in Twelve Libraries The purpose of this article is to identify the strengths of twelve North Ameri- can research libraries that have historic ARS collections; the twelve libraries are listed at the beginning of the article. The research began as a result of concerns voiced by the Academic Business Library Directors, 9 a group of fifty librarians that manage the largest academic business libraries in North America. The original opinion of the members was that the col- lections were highly duplicated (that is to say that we were all unnecessarily col- lecting reports from the same 800–1,000 companies). Therefore, a compilation of company names and dates would deter- mine how much overlap there was.10 An inventory of the combined collections could be used to locate needed reports for researchers when there were gaps in lo- cal collections. In addition, the inventory could be a guide to libraries that need to reduce their collection, as unique reports could be retained or offered to another library in the group. The list could also serve as a resource for a major digitization project; another goal of the group. Every librarian in the group was encouraged to send a list of the company names and dates in their collection; any format was acceptable. Yale University and New York Public Library did not have lists or the resources to compile lists, so their numbers in the charts below do not truly reflect the size of their collections. These reports were merged into a single data- base of nearly 64,000 records. To do the analysis, the list was then consolidated into one record for each company name with separate fields for each library’s holdings. The lack of au- thority control on the company names presented the major obstacle in the project; this was not surprising, as it has been a long-term problem for business librarians as documented in Stephen Bell’s article.11 The major problem was the inconsistent format for common words or abbreviations for company, corporation, and incorporated. So all these words were removed and the records were consoli- dated. Although there is some chance of two unaffiliated companies being merged together, the outcome for comparison sake is more valid. This merged database of approximately 38,000 records was used for this analysis. Results of the Analysis Collection Sizes The analysis revealed several interesting and some surprising facts. The first is the sheer size of the combined collection or number of companies. Together we have reports on 37,976 individual companies, even after merging to one record per company. The second surprise is that the largest collection, in terms of number of companies, is at Stanford University with 18,249 companies; and Baker Library at Harvard University’s collection, long thought to be the most comprehensive, is second in size with 11,504 companies. (This is just a comparison of number of companies; Harvard certainly has the deepest collections in years covered for each company.) Combining the collec- tions of Harvard and Stanford Universi- ties would produce a collection of 28,626 companies. Third in size is University of Western Ontario’s collection, with 4,513 companies. Combining Harvard, Stanford University, and University of Western Ontario would result in 32,151 compa- nies, or 85 percent of the total reports in the twelve libraries. Certainly these are the key collections in North America; the Annual Reports to Shareholders  529 future archive of printed reports relies on three universities, and one of them is Canadian. Next in importance are the collections at Columbia University and Purdue University: Columbia Univer- sity because of its historic depth, and Purdue University because of its more current reports. Adding these collec- tions would bring the size of a combined collection to 34,781, or over 91 percent of the total. The five libraries at Harvard University, Stanford University, Uni- versity of Western Ontario, Columbia University, and Purdue University form the major libraries with ARS collections and will be used in this article for bench- marking the other collections. Historic Depth of Collections The historic collections are of great- est interest for archiving as reports since 1996 are available in a number of commercial databases such as Mergent Online. Only a few libraries have col- lections in the 1800s. A look at the data shows, that as expected, Harvard has the deepest historical collection, but the Columbia University, New York Public Table 1 Total Number of Companies by library, Ranked by Size of Collection University/ library Number of Companies Submitted Total Number of Companies after Merge Stanford 23,422 18,249* Harvard 13,377 11,504 Western 4,826 4,513 Purdue 5,166 3,705 Columbia 4,674 3,491 MIT 2,407 2,295 Berkeley 2,257 1,328 Alabama 1,963 1,104 Penn 1,558 884 Cornell 507 435 NYPL 652 289 Yale 149 36 *Note: Stanford's collection included reports that were only retained for 5 years; these were removed in the figure in column 3. Table 2 Number of Companies at each library by Time Period, Ranked by Number of Pre-1900 Reports library Total Number of Companies Number of pre-1900 Companies Pre-1900 Overlap w/ Harvard Overlap Percent w/ Harvard Number of pre-1940 Reports Number of Post-1980 Reports Harvard 11,504 1,161 — — 6,079 1,454 Columbia 3,491 247 100 8.6% 2,781 0 NYPL 289 141 56 4.8% 262 3 MIT 2,295 117 7 0.6% 597 497 Stanford 18,249 33 14 1.2% 1,796 9,115 U of Alabama 1,104 10 2 0.2% 565 412 Purdue 3,705 9 4 0.3% 668 2,424 UC-Berkeley 1,328 0 0 0.0% 116 31 Western 4,513 2 1 0.1% 19 3,294 Cornell 435 0 0 0.0% 0 409 Penn 884 4 3 0.3% 243 81 Yale 36 1 0.0% 31 0 530  College & Research Libraries  November 2010 Library, and Massachusetts Institute of Technology (MIT) historic collections are noteworthy and very valuable. As the table above indicates, many libraries stopped collecting in the 1960s or 1970s when microform reports be- came available, so those that continued to collect printed reports have valuable collections even if their historic depth is not significant. Columbia University has a rich and deep collection, but stopped add- ing reports about 1966; New York Public Library’s collection focuses on dates prior to 1966 also. University of Alabama and Yale University stopped adding printed reports in 1979/1980, and University of California-Berkeley and University of Pennsylvania both stopped collecting in the mid 1980s. Cornell discarded their U.S. company reports but maintained their for- eign company reports from 1978 to 1985, making this a rather unique collection. Harvard University, Stanford University, University of Western Ontario, and Purdue University have strengths in the second half of the twentieth century. An example of the importance of reports after 1985 is that Proquest needed the Fortune 500 company reports between 1986 and 1995 to build the Proquest Historic Annual Report database. Only these four libraries among the twelve could supply these reports. The Myth of Overlap or Duplication of Collections The most surprising and important result of the analysis is that the collections are not heavily duplicated. The two largest collec- tions, Harvard University and Stanford University, only have 3,668 companies in common, and this figure only considers the names of the companies. If we compared names and dates, there would be even less overlap. Adding the University of Western Ontario to the Harvard/Stanford list brings the duplication number down to 236 com- panies. The overlap of companies among the big five collections is only thirty-three companies. Below is a comparison of each collection in the study with each of the five major collections. Conclusion So where do we stand today? There are only a few libraries that have strong holdings of printed annual reports collec- Table 3 Overlap of Companies with each Major Collection by library Number of Companies Columbia Harvard Purdue Stanford* Western Alabama 1,104 324 509 359 459 53 Berkeley 1,328 200 747 550 757 84 Columbia 3,491 — 1,484 483 1,287 133 Cornell 435 17 95 29 73 65 Harvard 11,504 1,484 — 1,608 3,668 371 MIT 2,295 99 166 91 84 37 NYPL 289 133 163 99 125 10 Penn 884 242 470 337 461 28 Purdue 3,705 483 1,608 — 1,546 302 Stanford* 18,249 1,287 3,668 1,546 — 644 Western 4,513 133 371 302 644 — Yale 36 36 32 36 32 8 *Eliminating all with Stanford holdings of "Latest 5 years". Annual Reports to Shareholders  531 tions. Looking at the data for the twelve libraries that submitted their holdings to the combined index, Harvard University, Stanford University, University of West- ern Ontario, Columbia University, and Purdue University form the major librar- ies with ARS collections: Four libraries have deep historic collections: Harvard stands out in this group, but Columbia University, New York Public Library, and MIT have important companies to supplement the Harvard collection. Late–twentieth-century collections (that is, post-1980), are available at Harvard, Stanford, Western Ontario, and Purdue. The overlap of these twelve collections is very low; this is contrary to the expecta- tions of the business librarians with these collections. In the recent past, several libraries that had historic collections have eliminated their reports because of the perception that their collection was not very unique. Before another collection is discarded, librarians should consult with the libraries with strong holdings. The database upon which this article is based can be used to analyze which reports should be preserved. The au- thor of this article is willing to prepare comparisons of collections and identify unique companies within a collection upon request. With a group effort, these valuable historic reports can be preserved for use by scholars. Notes 1. Judith R. Bernstein, “Corporate Annual Reports in Academic Business Libraries,” College & Research Libraries 47, no. 3 (1986). 2. Lisa O’Connor, “Managing Corporate Annual Reports,” in SPEC Kit, ed. Lee Anne George (Washington, D.C.: Association of Research Libraries, 2000). 3. Martha Jo Sani, “Does the Corporate Annual Report Have a Future?” Colorado Libraries 17 (1991). 4. Steven J. Bell, “Corporate Change: Impact on the Corporate Documents Collection,” Special Libraries 79, no. 4 (1988). 5. Cynthia L. Cronin-Kardon, Michael Halperin, and Di Su, “Annual Reports: Preserving and Disseminating a Source for Business History,” Acquisitions Librarian 19, no. 37/38 (2007). 6. Ibid. 7. Available at www.columbia.edu/cu/lweb/indiv/business/CorpReports.html. [Accessed 21 September 2010] 8. Columbia Broadcasting Service, Inc. 1950 Annual Report to Stockholders, New York: Columbia Broadcasting Systems, Inc., 1950. 9. Academic Business Library Directors’ Website. Available online at http://www.abld.org/ [Accessed 21 September 2010]. 10. Barbara Mento, “Academic Business Library Directors’ Meeting,” Journal of Business & Finance Librarianship 3, no. 1 (1997). 11. Bell, “Corporate Change: Impact on the Corporate Documents Collection.”