This item is only available as the following downloads:
CROWDING OUT IN PUBLIC RADIO BY MATTHEW CUTLER A Thesis Submitted to the Division of Social Sciences New College of Florida in partial fulfillment of the requirements for the degree Bachelor of Arts in Economics Under the sponsorship of Dr. Richard Coe Sarasota, Florida, May, 2013
ii Acknowledgements Thank you to Dr. Richard Coe and Dr. Duff Cooper, for the faculty support of my project And to, Walrus Research SRG, and the CPB for help in data acquisition And to, m y friends and f amily for everything else. Drut lives Mark Mathews
iii Table of Contents Table of Contents....iii Abstractiv 1. Introduction...1 2. Radio Present & Past5 3. Charitable Contributions and Crowding Out Theory.20 4. Empirical Link s...34 5. To Conclude58 Bibliography 64
iv ABSTRACT: CROWDING OUT IN PUBLIC RADIO Matthew Cutler New College of Florida 2013 The crowding out hypothesis was first empirically linked to the public radio market in 1989 by Bruce Kingma. Kingma found 13 cents of private funds to be crowded out for every extra government dollar contributed. Studies on the crowding out hypothesis continued ; questions regarding the v alidity of the explicit link between government and individuals became apparent. Further work makes clear that changes in fundraising expenditures are more directly linked to fluctuations in federal allocations to non profits, which in turn change private giving patterns. This thesis attempts to examine the impact to public radio stations by changes in public funding. Data from the years 2002 through 2011 was acquired from the CPB and tests were pe rformed with SAS and E Views Limitations in the available data forced an interpret ation of trends found solely within the financial information of 400 unique public radio stations (accounting for the total population of radio stations that receive public funding). Figures on CP B allocation, private donation revenues, and fundraising expenditures were used to determine that private giving rises alongside f undraising expenditures, but the resulting data supporting crowding in was questionable due to the lack of sufficient controls in the regression ___ __ ___ ____ ______________________ Dr. Richard Coe Division of Social Sciences
Cut ler 1 Chapter 1 | Introduction The FM band: 87.5 MHz to 108.0 MHz; most of it is ruled by commercials and pop music. And a lot of people like that. T here is a section reserved for more experimental formats, as well as public stations and religious broadcasters. It is within this reserved section of the band, in the range of the not for profit, educational, left of the dial, & low power broadcasters, which has entertained my attention. "S ee: commercial radio stations have a built in defusing process which is MAKE MONEY. What [community broadcasters] have is a group of dedicated sincere people who want to Do Good and Right." (Milam 19 88 ) 1 When a broadcaster's objective is not profit orient ed, there is a supposed freedom to create, express & document an unbiased vision. For a non profit radio station, there is however, a structure that needs to be satisfied. Costs are inherent in building and supporting a radio station and without the suppo rt structure that privatization brings, there is hope that private citizens find some value (equal to their pledge) in an inherently free to do system of audio transmission. Today, the United States and many countries around the world provide support for a public radio system. In the United States, federal monetary support for this system totals around 100 million dollars (but the Corporation for Public Broadcasting acts as a layer between the allocations and the stations). Businesses and non profits which find value in the public network sponsor programming through underwriting. Colleges and Universities also contribute monetarily to public radio. But it is the listeners, and record fairs, and volunteersand pledge drives, that provide the crux of monetary support, and the largest chunk of support at that. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 1 Community Radio Pioneer, Activist, and Writer. He founded the KRAB nebula of independent radio stations and continues to inspire many local, community broadcasters of the current age.
Cut ler 2 This thesis concerns the public radio network the network of stations that are the paying members of National Public Radio, American Public Media, Public Radio International & the Public Radio Exchange. NPR member stations are provided a mix of content from these larger distribution networks and from their own locally produced programming, and it is the interplay between the national and the regional that is so intriguing. The wide array of funding sourc es for public radio and other not for profit organizations has intrigued others as well. Where researchers, and the layman, are interested in the federal allocation structure of many non profits, work has been done in regards to the mixing, and subsequent relationships, of funding sources. One topic of inquiry is the interaction between the government and the individual. The individual is allowed to use their resources to support the non profit networks they believe to be of value. Consumers can decide thi s importance based on their monetary support. How important then, is the federal support of these non profit networks, and what role does federal allocation play in changing the support patterns in individuals? The theory of crowding out in non profits is that for a one dollar increase in government allocation, there will be some decrease in private donations to that non profit. There were theories that led up to the work of Bruce Kingma in 1989 that detailed the private consumer provision of public goods s uch as Abrams and Schmitz (1978), Warr (1982) and Roberts (1984), who estimated that crowding out happens at a rate of 28 cents, 1 dollar, and 1 dollar, respectively. 2 Kingma though was able to use a data set that matched household level data on private gi ving patterns with firm level data on radio station revenue sources. Kingma estimated that crowding out happened at a rate of decline of 13 cents per government dollar. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 2 Be rgstrom Blume, and Varian 1985
Cut ler 3 This study also references public radio. Using data from the Corporation for Public Br oadcasting, I test the empirical validity of Kingma's results, and the results of those that challenged Kingma's Brooks (2001), Payne (1998), Andreoni (2003), and others. Because of limitations in my access (and general restrictions on household data col lection), I cannot employ a data set like Kingma's for my study. I will instead attempt to track changes within the public radio market by accessing station level financial data. However, this data is not connected to measures of station outreach or market share (among other important but missing adjustments), and thus will be a somewhat crude interpretation of station level interactions. Further, this study attempts to deconstruct the crowding out hypothesis (with suggestions from the literature) by explo ring the linkage between government dollars and fundraising expenditures. Recent studies using expanded sets of data have begun to hypothesize that donations and allocations are not inherently related, and that instead, non profit organizations adjust fund raising efforts upon receiving an increase in allocated money. Changes in private donations then come about from decreased fundraising efforts. Where government allocations are found to negatively affect private giving or fundraising behavior, it may be o f value to government agencies to implement matching grant type programs which do not negatively inhibit the behaviors of private donors or non profit managers. This study focuses on the public radio market, but its implications for non profits are potenti ally widespread. Chapter 2 is an exploration of the foundations of public radio with a background on the breakdown of public funding, and the buildup of the educational broadcasting
Cut ler 4 network in which our current public radio system is rooted Chapter 3 addr esses the previous literature upon which the crowding out hypothesis is built, and how the theory of public goods relates, if at all, to public radio. Chapter 4 attempts to map the acquired data set to the evidential results in the previous chapter, to tes t the validity of the crowding out theory. Chapter 5 concludes and expands upon the results found.
Cut ler 5 Chapter 2 | Radio Present & Past The Federally Funded Public Radio Network Appropriations for public radio are funded federally, but are not distributed at the federal level. The Corporation for Public Broadcasting, born out of the Carnegie Commissions' 1967 report Public Television: a Program for Action, is a non profit organiza tion that requests federal dollars and distributes that money to a mix of recipients based on a statutory formula outlined in the 1967 report. There are direct grants to local television and radio stations. There are grants to television programming (PBS r eceives money directly from the CPB). There are radio programming grants which fund local radio stations to acquire national radio programming from producers such as NPR, Public Radio International, and American Public Media. There is a radio program fund where CPB money feeds directly into nationally distributed radio programming projects' such as StoryCorps and State of the Re:Union 3 There is a System Support fund which accounts for six percent of the CPB's total allocation, and is directed towards "pro jects and activities that will enhance public broadcasting". (47 U.S.C. 396(k)(3)(A)(i)(II)) Finally, as outlined in the Public Broadcasting Act of 1967, no more than five percent of the CPB's total allocation goes to support internal CPB administrative o perations 4 Considering the range of what the corporation funds, the percentage of the CPB budget directed toward radio specific ally is significantly lower than the allocation to television. Television receives around 300 million and radio receives around 100 million where the total amount of federal funds to the CPB in fiscal year 2013 is a little above !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 3 http://www.cpb.org/appropriation/ 4 Information for this brief outline on the division of the CPB allocation was sourced from http://www.cpb.org/appropriation/ and http://www.cpb.org/aboutcpb/financials/budget/
Cut ler 6 445 million dollars. System Support and CPB Administration both receive around 25 million dollars 5 This ratio of expenditures is not often contested, as television production and distribution requires more capital resources than radio. There is no significant debate surrounding this topic. Knowing that around 100 million dollars every year goes to support public radio is important though. Fiscal year 2010 saw 100 million dollars go to provide local stations on average, 10.9% of their total revenue 6 The largest portion of the average stations' revenue was attributed to Subscribers,' defined by the Corporation for Public Broadcasting as "revenue from membership, subscription contributions and friends groups." (CPB, 2012a) Between fiscal years 2000 and 2010, subscribers accounted for on average, a range of 32.0% and 36.4% of the revenue of each local station 7 On average, listener support to local stati ons provides more revenue than federally allocated dollars, though minority and native population stations may not be well represented in this statistic. Stations in these geographic areas receive less private contributions and a higher percentage of opera ting funds from the Corporation for Public Broadcasting. The revenue sources that remain consist of Colleges and Universities, Foundations, Business (from underwriting, etc.), State and Local Governments, Federal Grants and Contracts, and All Other. In 20 11, all non federal sources combined accounted for 89.2% of public radio's revenue. Business accounted for 19.2% of the total revenue; funds from state and private colleges made up 12.9% of the total; State and Local Governments contributed 4.2% of total r evenue; Foundations made up 8.0%, Federal Grants and Contracts made up a minimal 0.7%, and All Other sources made up !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 5 Corporation for Public Broadcasting FY 2013 Operating Budget 6 CPB, Public Broadcasting Revenue Fiscal Year 2010, T able 1 CPB, Public Broadcasting Revenue Fiscal Year 2010, T able 3. These figures exclude volunteer values.
Cut ler 7 the remaining 10.0%. The All Other category "i ncludes revenue from Subsidiaries, Special Fundraising Activities, Passive Income, Gains (Los ses) on sales of Assets or Securities, Endowment Revenue, and Capital Campaigns. 8 The majority of CPB funds that directly support public radio stations are allocated through the Community Service Grant (CSG) program. Of the money from the CPB that dire ctly funds radio stations, there is further division on how the appropriation must be used. There is a restricted and a discretionary portion of this CSG money. 75% of the CSG allocation is unrestricted' and 25% is restricted.' The restricted portion of the expenditures must be spent on n ational program production and acquisition. These funds must be used as specified in Section 16 396(k)(3)(A)(iii) of the Communica tions Act of 1934 (47 U.S.C. ¤ 396(k)(3)(A)(iii)), which provides that the fu nds are solely to be used for acquiring or producing programming that is to be !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 8 CPB, Public Broadcasting Revenue FY 2011 34.8% 19.2% 12.9% 10.1% 10.0% 8.0% 4.2% 0.7% Figure 2.1: Public Radio Revenue Sources FY 2011 6 Subscribers Business State and Private Colleges CPB All Other Foundations State and Local Governments
Cut ler 8 distributed nationally and is designed to serve th e needs of a national audience." (CPB 2012b) These restricted funds must be limited to ( 1) Production (2) Program and Content Acquisition (3) Distribution & (4) Promotion 9 While the Corporation for Public Broadcasting does not wholly and directly support say, National Public Radio or American Public Media, there is a portion of funds delivered to member stations that is explicit ly requested to subsidize these non commercial content providers. Both federal support and private contributions for nationally syndicated public content are directed through local member stations. Even where a majority of national content funding is fed through local public radio stations, the knowledge of how national content is financially supported is not overtly public. In a 1999 public radio trend report, The Public Radio Tracking Study by George Bailey, participants in an Arbitron study of public r adio listening were re contacted and questioned about topics including Fund Drive Annoyance, Personal Importance, and Government Support Beliefs. In reference to government support beliefs, there was a general trend noting that non contributors (compared t o contributors) were more in agreement with the statement that public radio is financially supported by government tax dollars. This might show that persons who decide to contribute realize in some way that their individual contribution is more powerful to the public radio network (in terms of the percentage of private dollars versus government dollars and the support these sources provide) than the supposed benefit a non giver values their contribution at. Possibly, where individuals believe the government 's support of the public radio network !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! # Corporation for Public Broadcasting FY2013 Radio Community Service Grant General Provisions and Eligibility Criteria
Cut ler 9 is strong, fewer private donations are made. This evidence is not conclusive, as there are multiple factors such as personal importance and income to consider, but an interesting question is brought up: how important is the consideration of federal funding in a private donor's choice of contributing to public radio? Examining how government funding crowds out private contributions in non profits has been the focus of work in economics dating back around 30 years, with early theory on the voluntary provision of pu blic goods dating back to Olson's 1965 book, The Logic of Collective Action: Public Goods and the Theory of Groups. Though public radio cannot be modeled explicitly by public goods theory, there are some theoretical implications to be drawn on, as explaine d in Chapters 3 and 4 of this work. The Foundations of Educational Broadcasting A federally funded public radio network is the topic of some dissenting opinions, for reasons about the validity of the cultural value a public media network brings, the edu cational content it provides, and the ability of private structures to provide the content that is cherished by some listeners and critics. The fight for a public media network began in the early 20 th century. The Corporation for Public Broadcasting was n ot the beginning of non commercial radio in the United States. Pre dating the CPB, the Federal Communications Commission was established with the passing of the Communications Act of 1934. The FCC was the first permanen t commission that would replace previ ous state regulations for radio, telephone and telegraph industries with national regulations. Many pre FCC experimental radio stations were founded and maintained by educational and religious
Cut ler 10 institutions, and with the Communications Act of 1934, many "ed ucators, churchmen, and labor leaders came forward, stressing radio's educational and cultural potential," in the fear that this newly organized broadcast spectrum and the commercialization of the industry may push them out. (Witherspoon et al. 2000 ) An amendment to the 1934 bill was brought forward by S en ators Robert F. Wagner (D N.Y.) & Henry Hatfield (R W.V.) to "reserve and allocate one fourth of all radio broadcasting facili ties to nonprofit stations." (Witherspoon et al. 2000 ) Cultural significance' was a factor suppor ters pushed. Wagner's proposal was rather extensive though, calling for a withdrawal of all existing licenses, "and reallocating frequencies, power and operating hours for all stations in 90 days." The frequency spectrum was also supposed to be arranged s uch that commercial and nonprofit stations were comparably distributed. The Wagner Hatfield amendment proposal was defeated 42 to 23, but Senator Clarence Dill "included a section in the  act requiring the FCC to study assigning channels to nonprofit organizations." The FCC released its report to congress in January of 1935 recommending against any fixed percentage amendment. The struggle for a dedicated spectrum assignment continued. We can look even further back to commissions established before t he FCC. A variety of organizations attempted to consolidate American public broadcasters beginning in the late 1920's. Educational, experimental radio stations, based mainly out of universities, formed the base of what transformed into public broadcasting. In a sense, public radio was first realized as a group of educational institutions whose individual purposes, to experiment and educate with this new technology, were hoped to be expanded on, and nationally connected, to create a powerful network of educa tional
Cut ler 11 content producers and distributors. The field of radio was expanding and as commented on by Herbert Hoover (then Secretary of Commerce) at the Third National Radio Conference in 1924, "radio has passed from a field of adventure to that of a public u tility." 10 At the time, the Department of Commerce was largely in charge of radio regulation. It was not until the Radio Act of 1927 was signed into law that a radio specific governing body was created the Federal Radio Commission (FRC). The FRC was tasked with allocating station licenses. By 1929, new frequencies and power levels for existing stations were arranged. The FRC established General Order 40, which designated a clear channel network comprised of 40 separate 50,000 watt national stations. NBC an d CBS were two of the original commercial broadcasters and together in 1927 they oversaw 6.4% of broadcasting stations. In 1931 (after the FRC had reallocated spectrum assignments) NBS and CBS effectively broadcast to 70% of the market and operated 37 of t he 40 clear channel frequencies 11 This rapid expansion was due in part to the contestability of radio frequencies and the well capitalized' nature of private broadcasters. Stations were left to compete with one another for transmission hours and eventuall y, frequencies. Commercial interests secured a large portion of the market. Robert McChesney, a researcher and professor who has thoroughly examined the political economy of media development in the United States, helped to assess the conflict and critic ism that came with the regulation of radio and the commissions involved. He addresses the adoption of our current media system (and the roots of such !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! $% Recommendations for Regulation of Radio Adopted by the Third National Radio Conference October 6 10, 1924, issued by the Government Printing Office in Washington, D.C. for the Department of Commerce [C1.2:R11/924]. 11 Clear channel frequencies' refers to: limiting one frequency to one station, not the private American media company. McChesney 2008.
Cut ler 12 adoption) and explains that the American media system was in fact contested in its formative years. In Rob ert McChesney's 2008 collection of journal articles, commissioned works, and political analysis, The Political Economy of Media: enduring issues, emerging dilemmas he explains that much in line with Herbert Hoover's sentiments regarding the public utili ty of broadcasting an opposition movement to the Federal Radio Commission that accused the FRC of favoring commercial broadcasters thought that "the airwaves should be regarded as a public resource and broadcasting as a public utility ( McChesney 2008 ) Various associations of broadcasters and other non broadcasters such as the American Civil Liberties Union (ACLU) were fighting for non commercial broadcasting rights leading up to the Communications Act of 1934 and maintained t he idea that commercial radio was missing minority opinion s and in suit, many cultural, educational, and public affairs programs would not air on an advertiser subsidized radio. The FRC was much to blame explains McChesney, as the general attitude from the FRC in 1929 was that ca pitalist broadcasters upheld a general public service,' as "their quest for profit would motivate them to produce whatever programming the market desired." Non profit stations not raising revenues from advertising "were termed propaganda' stations" those more interested in spreading their viewpoint than in satisfying audience needs." (McChesney 2008) With a limited number of frequencies to distribute and an increasing number of licensing applications, general public service' stations were overwhelmi ngly favored by the acting commission. It was only after commercial stations began to draw listeners in the late 20's and early 30's while educational stations failed to maintain listenership for a variety of internal and external reasons that organi zation by educational broadcasters became
Cut ler 13 somewhat necessary. Though the Association of College and University Broadcasting Stations (ACUBS) formed in 1925 to rally in support of educational broadcasting, it did not see a dedicated following, and these ear ly years saw the dominance of commercial broadcasting steadily increase. It was with ACUBS that the establishment of a dedicated frequency range for educational broadcasting was first pushed. A national headquarters for educational radio was pushed. But no ne of this would succeed under ACUBS. As the industry advanced, the committee re established itself as the National Association of Educational Broadcasters (NAEB) in 1934 and would see some of their ideas for an educational radio network come to fruition i n the late 30's. What lead up to 1934's communication act was mostly a struggle between the reality of ever strengthening networks like NBC & CBS, and the hopes of some various education based committees who analyzed and evaluated the corporate structure that the industry was starting to resemble. A major contributor to these educational broadcasting studies was the Payne Study and Experiment Fund. This fund promoted two campaigns of its own the National Committee on Education by Radio (NCER) & the Vent ura Free Press and subsidized other efforts of research into the changing radio industry. The Payne Fund was a non profit organization active in the late 20's and beyond who offered financial support to research on the interaction between the media and e ducation, among other things. The fund wanted to ensure that content with educational value would survive free of influence from advertising purposes. In 1929 the Secretary of the Interior Ray Wilbur called for a meeting addressing educational broadcasti ng (the meeting was paid for by the Payne Fund.) From this meeting, the Advisory Committee on Education by Radio (ACER) was formed with the
Cut ler 14 goal of studying "the crisis in educational broadcasting" and reporting on those findings to Secretary of Interior W ilbur. This committee too was subsidized by the Payne Fund, but also from the Carnegie Corporation. ACER was made up of representatives from nine different associations, ranging from "education, broadcasting, manufacturing, government, and public" sources (Witherspoon et al. 2000 ) ACUBS representatives would sit on this committee, as would representatives from the Natio nal Educational Association (NEA) and the National Association of State Colleges, among others Armstrong Perry worked on the committee full time as he interviewed broadcasters and educators across the United States. According to McChesney, Armstrong Perry 's investigation into the industry radicalized' him. Perry was supposed to address the possibility of implementing educational broadcasts within the commercial system, but after talking with many educators and broadcasters, "he became convinced that the o nly hope for education on radio was to protect these [university] stations" because any "commitment of NBS and CBS to provide free airtime and facilities was evaporating." (McChe sney 2008 ) The state and university network of broadcasters had fallen to the commercial friendly regulations of the FRC. ACER recommended to Ray Wilbur the reservation of channels for educational transmissions; "establishing an educational radio division in the Interior Department's Office of Education; establishing a [more perma nent] advisory committee of educators, commercial broadcasters, and the public; securing funding for education programs; and alerting the President and the Federal Radio Commission of the importance of educational programming to the American Public." (Witherspoon et al. 2000 )
Cut ler 15 A new radio division was in fact created in the wake of these recommendations, and Perry oversaw this new office. Perry now worked under the Commissioner of Education, William John Cooper, and was able to get Cooper to call the October 1931 meeting that es tablished the National Committee on Education by Radio (NCER). This committee was focused on creating an educational broadcasting system separate from the commercial radio industry, opposed to earlier efforts to find an accommodating space for educational programming within the commercial marketplace. In 1931, stations licenses were only valid for three months at a time. This meant that the FRC was able to evaluate each station's viability and allow new broadcasters to take over where the FRC saw fit. As radio as an industry developed and the medium became more popular, private entities filled the market (usually NBC or CBS affiliated). One job of the NCER was to lobby for these educational station licenses; hopefully this would allow educational programs to develop without the concern of devoting time and funding to licensing issues. More and more stations continued to be turned over to commercial means. Also created out of the ACER recommendations was the National Advisory Counci l on Radio in Education, an organization separate from the NCER in structure and funding. NACRE was founded by the Carnegie Corporation, and funded by the Rockefeller Foundation. This committee worked with NBC to try and create educational content for the commercial broadcaster, but received very little support from the rest of the educational broadcasting community. The Rockefeller foundation discontinued funding in 1938, marking the end of NACRE. While both NACRE and NCER were working for education in radio, the two clashed on t echnique and in no way worked
Cut ler 16 cooperatively. This may have caused confusion within the community, but the strong political lobbying of the commercial entities through the National Association of Broadcasters (NAB) was difficult to counteract, with or witho ut internal hindrances. NAB was founded along with General Order 40, and would go on to represent and lobby for most of the radio industry after the passing of the Communications Act of 1934. Even with all of the efforts of the educational community for a public broadcasting structure, the Communications Act of 1934 would go on to solidify the commercial infrastructure of the industry. As the FRC was re built into the FCC, many of the considerations for non commercial media stayed the same. Educational bro adcasters were allocated no frequencies. I t was also at this time that ACUBS would re evaluate its mission and name; the National Association of Educational Broadcasters was established and continued to work towards a form of non commercial radio. NAEB and others pursued broadcast reform and in 1938, the FCC outlined a new classification of high frequency Noncommercial Educational Broadcast Stations. The classification of these stations did not include any channel reservations at first. The Federal Radio Ed ucation Committee (funded by the General Education Fund who received money from the Rockefellers) advised the FCC in 1939 that "cooperating with broadcasters in promoting education through the medium of radio'is a mutual necessity,'" after studying the is sue since 1935. In 1940, five high frequency channels were specified for Noncommercial Educational Broadcast Station use. Still, financial support for these stations was limited. By 1945, six of these stations were fully operating, "construction permits fo r six others were outstanding, and 22 applications for new stations were on file." (FCC 1945) Interest was increasing, and the FCC allocated 20 FM channels in the band 88 to 92 megacycles
Cut ler 17 (where the commercial frequencies occupy the channels associated wit h 92 to 108 megacycles) in order to accommodate this growing need 12 In 1949, the FCC acknowledged that the amount of Noncommercial Educational stations grew from eight to 22 between 1947 and 1948, but also acknowledged that the interest from educational es tablishments has been somewhat deterred due to the "lack of sufficient appropriations by school systems" in the post war era. (FCC 1948) It was also in 1948 that the FCC first issued regulation for 10 watt low power stations (though not officially LP' or Low Power') as a way for educational broadcasters to adapt to tight financial situations as these lower wattage stations required a smaller financial commitment. The Ford Foundation provided needed help that kept educational radio afloat through to the la te 60's 13 The industry as a whole continued expanding, and by 1967 there were 412 educational stations broadcasting at various wattage levels (and to various degrees of success). It was not until this year that federal support was finally legislated by the Public Broadcasting Act. National Public Radio soon followed. Evidently, there were some efforts to acquire educational channel allocations following the creation of the FCC, but commercial networks became and remained the main focus of the industry. Many thought this was for good reason; the commercial structure brought upon a consolidation' of the industry and a transformation of "the character and quality of programs." It also secured mass investment in the industry with large amounts of revenue fr om advertising efforts 14 What McChesney iterates is that the commercially driven American broadcasting system is the way it is today because of the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 12 FCC Eleventh Annual Report, 1945. Pg. 22 13 The Ford Foundation started supporting public broadcasting in 1951 through to the eventual establishment of the CPB. NPR: The Trials and Triumphs of National Public Radio by Michael P. McCauley. Pg. 6 14 A History of Public Broadcasting pg. 9. Quoting historian Charles Siepmann.
Cut ler 18 failure of educational broadcasting in the mid 1930's and because the FCC set a precedent for the continuing dominance of American commercial broadcasters to this day, and not because the private radio sector is a product of some organic, American, economic system. The critiques, he explains, from those who hoped to reform broadcasting were all very similar. (1) Where radio was funded by commercial entities, how free could radio be in reporting on "big business and the status quo (commercial stations could not provide the critical reporting that democracy needs)? (2) Where programming was advertiser driven, how e ffective could programs be in providing "educational, cultural, or controversial fare" when for profit content would be produced at minimal cost and made to cater to a majority audience? (3) Where public insight was blind to the rise of the commercial syst em and "where even Congress had played almost no role," what about the commercial broadcasting system represented a democracy? (McChesney 2008 ) Members of the reform movement felt that radio should have been established as a public utility, for the public good, but they were instead handed a system built around private interests. Public radio today is strong, but of course, is at odds with the com mercial spectrum. The continued government support of this alternative to commercial media is embraced by some and criticized by others, but yet, the single greatest source of revenue for public radio stations is from listeners. This thesis explores the fi nancial interaction between the government, individual donors, and the fundraising efforts of public radio station over the time period of 2002 to 2011 to explore the financial strength of the public radio market, the idea of public radio as a public utili ty, and how central the never ending pledge drives are to public radio's financial security.
Cut ler 19 Chapter 3 | Charitable Contributions and Crowding Out Theory Literature on the Effect of Government Subsidies to Charitable Giving The breadth of research on cha ritable giving to nonprofits is extensive, and the main exploration of these researchers is determining what level government funding crowds out private contributions to a non profit group such as a public radio station. Foundational work on the crowding out effect was done in the 1980's, and this work led to a much referenced 1989 study by Kingma. Early work on the crowding out model was "derived from the assumption that individuals see their own contribution as a perfect substitute for dollars given by the government." (Andreoni & Payne 2008) But, public goods theory can be drawn on to explain the presence of free riders in a market where a good is equally provided to all. Early perfect substitution crowding out theory would later be characterized as pu rely altruistic,' and naturally unobservable, by Andreoni (1988); a theory of warm glow' giving was subsequently proposed. Kingma (1989) went on to note that increases in government contribution crowd out private donors to some degree, and Kingma's work w as signified a landmark in the field of non profit, private contribution modeling, as he was able to empirically analyze contributions to public radio stations where work was previously limited empirically. For every extra dollar in government funds to pub lic stations, the total privately donated amount fell by about 13 cents. Kingma's study found that as the government increased funding by 10,000 dollars, individual agents decreased their donation amount by 15 cents. Where the average member count of publi c radio stations was 9000 members, a total crowd out of 1,350 dollars was calculated as the crowd out total resulting from the 10,000 dollar government subsidy increase.
Cut ler 20 Questioning Kingma Kingma's 1989 study remains widely cited because Kingma received s tatistically significant results of crowding out and was able to achieve that by matching household level contributions data with station level revenue data; such detail in the data set had not previously been explored. (Manzoor and Straub 2005 ) Manzoor and Straub (2005) wanted to re examine Kingma's study by replicating his methodology on a newer dataset. Kingma's estimates on the original 1986 dataset were replicated by Manzoor and Straub, who then proceeded to apply that methodology to a 1996 dataset. They found that the Kingma's estimation was "not robust to the use of newer data from 1996," and interpreted that this was possibly due to misspecifications in Kingma's original model. (Manzoor and Stra ub 2005 ) The authors further stated that definitive evidence' for the crowding out of private contributions was not likely, due to the available data "the ideal dataset would match firm level data from a panel of nonprofits to household level data on contributions to those same nonprofits over time." (Manzoor and Straub 2005 ) Household level panel data on contributions to public radio has not been available in recent years 15 though some research has bee n done on station level panel data (Payne 1998; Khanna & Sandler 2000; Ribar & Wilhelm). Panel data on household contributions matched with the station level panel data is not available. Audience Research Analysis' Audience 98 report did draw on data colle cted from individuals about their contribution behavior in the past, but new matching data has not been collected. Brooks (2000) noted that Kingma's study was found to be consistent with some other work in the area of non profit funding, though when speci fying Kingma's work in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 15 A financial director at the Radio Research Consortium explained to me in a phone conversation that collecting such data would be an extreme undertaking, and such data collection will not likely be performed again due to human constraints (strenuous amount s of time and energy).
Cut ler 21 the context of arts and culture non profits (art and history museums, orchestras, public radio stations), the results are somewhat inconclusive. Brooks (2000) examined a selection of previous work done in the field (with regards to a ll non profit research, not just arts and culture') and was able to generalize that the crowding out of private gifts due to increases in government allocations is either small or nonexistent, though the variety of results from studie s are not easily gene ralizable. Brooks states that "if a broad pattern is visible, it would not be in the direction of crowding in." (Brooks 200 0 ) Crowding in would be a situation where private giving to a non profit increases, when government funding to that non profit als o increases. The author's personal research further shows no evidence of crowding in to any non profit sub sets, a small but significant level of crowding out in social human service non profits, and no significant interactions in arts and culture non prof its. One reason for Brooks' division and classification of non profit sectors is due to the different characteristics of the donor groups; donors to public radio and donors to the opera can be characterized as high income, highly educated consumers. "Fundi ng for services such as the arts and higher educationare generally demand inelasticsome of those who give to these subsectors [might] do so more for social motives than for reasons of need." (Brooks 2000) Brooks (2001) explains that the impact direct g overnment funding of public radio has on giving behavior is an effect of either leveraging or displacing private giving 16 He also adds to the conversation what indirect funding non profits may receive due to changes in income tax rates, specifically addres sing changes in the price of giving. The !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 16 Direct public funding is money from the CPB, state and local governments, & other federal grants. Indirect government funding takes the form of special tax treatment that charitable contributors receive.
Cut ler 22 price of giving is the cost of a taxpayer to give one additional dollar to a tax deductable non profit, one minus the marginal tax rate. However, such an individual tax payer would need to itemize their tax deductio ns to receive this benefit; high income donors may be affected by changes in the price of giving in response to changes in the marginal tax rate, though lower income donors generally do not itemize their deductions. Still, as incomes rise, it is suggested that charitable giving amounts rise, and price of giving effects should be considered in policy initiatives. The topic of displacement or leveraging is a question of intention behind the crowding out effect. Are private donors "encouraged to give more wh en the government contributes, due to reputational effects, better information about recipients, and matching fund arrangements," or do donors "perceive less need' when the government funds a nonprofit organization, and therefore donate less..?" (Brooks 2 001) The author looked at CPB records for 104 radio stations for the years 1990 to 1996. Brooks' study took into account all non profit stations with assets more than 10 million dollars, and included a probability sample of smaller stations. The author als o included data on state tax rates and per capita personal income. Brooks concluded that the level in which government contributions crowds out private donations is not constant as detailed by Figure 3. 1. At low levels of public funding, private giving is crowded in. This effect changes as government subsidies increase, with high level government funding crowding out private giving. Brooks also concludes that higher state tax rates correspond with more donations per capita to public radio non profits. No n profits receive heightened contributions whe re taxes are higher as contributors forego some amount of taxes, and Brooks explains that
Cut ler 23 the government contributes indirectly to this as paying for part of that gift in those foregone taxes. In a "meta analys is of the economic analyses on this topic Steinberg (1990) found that on average, a 10 percent increase in the tax rate leads to about a 12 percent increase in charitable donations. (Brooks 2001) Brooks explains the correlation is due to research practices of holding income constant within a stud y such that the noted rise in donations actually comes from differences between tax brackets a non surprising trend. (Brooks 2001) Studies explaining that higher taxes lead to greater donations (Brooks i ncluded), are a product of single year interpretations of constant income tax levels which simply suggest that government indirectly pays a larger part of the donation due to the differing tax levels rather than real changes in tax policy. 17 Figure 3.1 : Cro wding In or Crowding Out of Private Donations at Different Levels of Government Support Per Capita !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 17 The government will forego collecting 35% of each contributed dollar when a tax payer with a marginal tax rate of 35% spends that dollar contributing to charity, and only 15% of a dollar that is contributed by a person who received a 15% marginal tax rate. A higher tax rates means a greater indirect contribution from the government.
Cut ler 24 Warm Glow: Public Radio Public Good Kingma and McClelland (1995) tried to model the charitable giving behavior under the restrictions of a public goods hypothesis. They showed that in the field of public radio, charitable contributions were positively affected by level of use of radio (whose output can be described as a pure public good.) This suggested that the provision of public radio cannot be modeled as a public good. Economists have spent considerable effort explaining the impetus for rational agents to make charitable contributions. Ea rly theory that cha ritable contributions supporting the output of a non profit could be modeled as a pure public good was refuted by Sugden (1982, 1985) and Andreoni (1989), "who showed that with a significant amount of donors, free riding would drive cont ributions to minimal levels." In addition, regarding the neutrality of contributions and the idea that government funding would crowd out contributions "dollar for dollar" implied by a pure public good model only Roberts (1984) showed a dollar for dolla r crowd out while many other sources have shown magnitudes less than a dollar for dollar effect. (Steinberg 1991) These sources support ed an impurely altruist ic model of charitable contributions. Impure altruism states that donors receive personal satisfa ction from making a charitable contribution, on top of already receiving the benefit derived from the public good. Kingma and McClelland's (1995) original research first tested for evidence of crowding out in public radio, and further tests explaining a le ss than dollar for dollar crowd out. The authors used a data set which consisted of radio listening, levels of charitable giving, income before tax, age, level of education, and other individual characteristics provided for a 1986 Arbitron survey and a 198 7 Arbitron re contact survey. The research showed that an individual donor (household) will react to a
Cut ler 25 $10,000 increase in government expenditures with a 15 to 19 cent decrease in contribution. This conclusion was similar to the conclusion found in Kingma' s original 1989 study. The authors suggested explanations for giving patterns, such that it was also estimated that a $10,000 dollar increase in private income increases a household's contribution from $3.57 to $6.82. They suggested that donors treat chang es in government funding differently from changes in the contribution of others, and they also suggested that if a donor's listening to public radio increases by 15 minutes, that individual's contribution will increase by around 50 cents. It was concluded that contributions are positively affected by the amount of public radio listenership and also that a government allocation to a public good decreases private charity and the "rate of decrease is inconsistent with the pure public good [hypothesis.] (Kingma and McClelland 1995) James Andreoni (1990) showed that a purely altruistic model of charitable giving is not easily generalized and that the correct model of charitable giving was one of Impure Altruism.' Two extremes detailing altruism were exa mined. Theoretically, an individual could be purely altruistic : where the individual cares for no form of reimbursement for the private gift or purely egoistic : where the individual is motivated to give only by warm glow' motivatio ns. Where factors of wa rm glow were present but not fully responsible for gift giving an individual could be described as impurely altruistic. Andreoni conclude d that individuals who make donations to privately provided public goods "may not only gain utility from increasing its total supply, but they may also gain utility from the act of giving." (Andreoni 1990) However, the public goods model cannot account for charitable contributions to public radio. Where a simple public goods
Cut ler 26 model will explain that "total provision of the public good is independent of the distribution of income among contributors, government provision completely crowds out private provision, and subsidies are neutral," empirical evidence provided by Andreoni (1990) sho wed th at a model of impure altruism was more consistent with ch aritable giving. This meant is that individuals were "not indifferent between gifts made by [them] and gifts made by other individuals or the government." (Andreoni 1990) Private donors are not indifferent between funds contributed by themselves and government entities. In the United States, income taxes jump up with increases in income. It has been shown that weal thier listeners to public radio donate in greater amounts. Ceteris Paribus, wealthier individuals who contribute the greatest amount of tax dollars would contribute less by means of private donations because those wealthy private contributors are already c ontributing the greatest amount to the tax system which subsidizes public radio. But there are other factors to consider. Tax write offs may be obtained only by individuals with high incomes, the marginal price of giving in wealthy individuals is less than non wealthy donors, and demand inelasticity in certain sub sectors of non profits all contribute to specific giving patterns of public radio contributors. Levels of income and education have already been shown to be reliable indicators of whether a public radio listener donates, and the level at which a donation is made. Andreoni claimed that finding a way to distribute preferential tax treatment to differently altruistic people may actually help to increase total financial support of non profits. This wa s expanded on in later work (Brooks 2001) that examined changes in the price of giving due to changes in t he tax structure. Andreoni showed that income
Cut ler 27 elasticity rose as i ncomes grew to $10 0,000 dollars, while altruism fell Therefore, the $50,000 to 100, 000 income group ranked "as the least altruistic." It was suggested that because of this, cutting taxes in the highest income group ($100,000 or more), while incr easing taxes in the 50,000 to 100,000 dollar income gr oup may increase private giving. C utting taxes to this group, and raising taxes on all groups below, "may actually reduce total charity." Classic V ersus Fundraising Crowd Out Contributions by Payne & Andreoni Payne (1998) added to the empirical literature on the crowding out effect by utilizing a panel data set of 430 non profit organizations over a 10 year period, 1982 1992 (though not any public r adio stations). The author linked private donations to non profit firms with the government grants those firms received while controlling fo r firm heterogeneity, macro shocks, political and economic effects, and potential endogeneity. Using state level political and economic indicators matched with donations collected by non profit organizations from those states, some political and economic e ffects could be controlled for. Crowd out was found to be essentially zero,' using an ordinary least squares specification. Using a two stage least squares specification, Payne found a partial crowd out of around 50 cents per government dollar. These find ings are varied. Payne concluded that crowd out in the social services sector of non profit organizations existed but was not complete; p artial crowd out was a possibility The author confirmed the robustness of both OLS and 2SLS specifications, but prefer red the 2SLS conclusion. Payne used mostly social service non profits as data sources in the study, which helped to
Cut ler 28 provide differentiation between various non profit entities, rather than trying to cluster various non profit organizations into one study. One foundational element of this research that later helped to characterize empirical non profit crowd out research was how Payne used a data set much richer' than previously used, Payne noted that further research should be done on firm level interaction s, exploring the role of the non profit entity in changes of a donor's behavior. Payne (2009) analyzed a breadth of previous literature on "the effects of government programs on individuals and organizations," which familiarized readers with the empirical estimation issues that were somewhat commonplace in examinations of the topic. (Payne 2009 ) Some recent investigations had focused on an attempt to comprehensively model the public radio and nonprofit market. They examined how government funding changes affected firm level fundraising tactics which in turn affected contributions to nonprofits, rather than a direct crowding out of contributions by increases in federal subsidy; this is one topic. Another was issues of endogene ity in the model, where government funding may have been related to the dependent variable of state/local funding, private donations, or private spending "attributable to the political process," such that "government officials awarding the grants to charit ies are elected by the individuals who donate." (Payne 2009) This issue followed Payne's (1998) previous research, and was addressed in earlier works by Steinberg (1987) and McClelland (1989). In the past, 2 stage least squares specifications were used to address this. Public Radio was not specifically addressed, though these empirical modeling specifications are important to consider in empirical applications.
Cut ler 29 Issues of indirect crowding out and the limited use of panel data was addressed in a 2011 paper by Andreoni and Payne titled: Is crowding out due entirely to fundraising? Evidence from a panel of charities A distinction was made between a classic crowd out' where consumers felt that an increased government subsidy replaced their contribution amount, and a fundraising crowd out,' where charities reduced their fundraising efforts upon receiving a government grant. The research investigated the effect of a government grant to non profit fundraising efforts. The study revealed that the decreased fundraising of the charities was "responsible for essentially all of the crowding out," and that private contributio ns may not be directly linked to changes in a tax based government subsidy. (Andreoni & Payne 2011) Crowding out was found to be significant at approximately 76%. Depending on which specification of organizations Andreoni & Payne included, classic crowding out could account for a range of 30% to a slight crowd in effect, where fundraising crowd out attributed to a range of 70% to over 100% of crowding out. (Vesterland 2011) In the specifications that Andreoni & Payne preferred, a $1000 grant would r esult in a classic crowd in of $45, a reduction in fundraising by $137, and a fundraising crowd out of $772. (Vesterland 2011) With these specifications, a non profit would see its net revenue rise $410 dollars from a $1000 dollar grant. Crowding out was t hen said to be a result of grant sensitive changes in fundraising practices of charities who may not be net revenue maximizers.' Within a context of non profit work, where benchmark fundraising goals are set rather than a goal of increased profit returns, fundraising crowd out' may make sense. Andreoni & Payne (2011) pointed out that fundraising efficiency within the non profit industry does not follow typical market efficiency standards, as something like
Cut ler 30 high spending (in order to maximize revenue) on f undraising campaigns may trigger poor quality ratings by various non profit watch groups. Andreoni & Payne (2011) outlined the importance of work in this field, namely that understanding the magnitude and causes of crowd out in non profits can increase n on profit efficiency. The efficiency of non profit organizations was a sustained focus of both author's previous research. Fund raising efforts were distinctly addressed. Specific limitations in previous econometric calculations (of both these authors and others) of crowding out were mentioned, examples being that charities were suspected to respond to a grant by reducing fundraising efforts, and that private donors are not suspected to have total knowledge of changes in government funding to non profits. A ndreoni and Payne characterized crowding out as a potential "hidden cost" of government subsidies. Andreoni and Payne also wanted to further research the behavioral models of non profits (and specifically charities) to see whether grant funded non profits are revenue maximizers. They note that research also continued into the behavioral models of non profit donors and the viability of the theory of warm glow' giving. Finally, where research could explain if "crowding out is in fact due to endogenous respon ses of the charity, it [could expand] the policy tools available to a government wishing to maximize the benefits of the tax dollars spent." (Andreoni & Payne 2011) Andreoni & Payne (2011) further emphasized strong reasons to believe that fund raisers an d charity managers are not net revenue maximizers, noting results from their study, data from older studies, and other social factors and fundraising methodologies. Their study followed three interactions: (1) Effects of government funding on private donat ions. (2) Effects of fundraising expenditures on private donations. (3) Effects of
Cut ler 31 government funding on fundraising expenditures. They found that the first effect was significant (and was similar to Payne's 1998 research) and reported various instrumental variable estimations that showed a total crowd out of private donations to be around 76% (~76 cents per dollar for government grants received). Though this was less than dollar for dollar, the effect was large. The 2 nd effect showed that a marginal dollar spent on fundraising yields between five and six dollars in new donations on average. This yield is typical of fund raising campaigns. After examining the 3 rd effect, Andreoni and Payne suggested that fundraising expenditures were reduced by 14 cents per dollar for government grants received. The researcher's provided calculations to differentiate between the total, direct, and indirect crowd out to suggest the possible causes of this private response by combining the results from the three effects. Agai n, the total crowd out of private dollars could be found by solely looking at effect (1), but it was emphasized that this does not explain the observed effect of government subsidy. The direct (classic') and indirect (due to fund raising') crowding out o f donations was calculated by combining "within column coefficients [which produced] a sample of estimates," that represented the interactions of changed donations by grants, changed donations by fundraising, and changed fund raising by grants. (Andreoni & Payne 2011) From this, the authors showed that most of the crowd out was likely attributed to decreased fund raising. One possible situation exemplifying Andreoni and Payne's conclusion explained that "a $10,000 grant, for instance, [will reduce] fundrais ing expenses by $1410, which in turn [will reduce] donations by $7570. Accounting for the 1 410 dollars in fundraising savings [will reduce] the estimate of crowding out to 66% ." (Andreoni & Payne 2011)
Cut ler 32 The research by Payne and Andreoni referenced a pan el data set of non profit entities, and no public radio figures were included in their study. Charity based non profits were the focus. While considering this, the theoretical conclusion from the study was that in the case where charities maintained the fu nd raising practices (in the position of receiving some additional government grant) of pre grant efforts, donations would have risen by the full amount of such grant. Andreoni and Payne do consider various specifications to their data set with a focus on the types of charities included/excluded, but find little significant differences when changing these specifications. A wide spectrum of non profit organizations was considered and the conclusion was maintained. For this reason (among others), the study ma y be applicable to public radio fund raising practices. On the Horizon The next chapter performs empirical checks on the viability of the theory of crowd out specific to the public radio industry. I will look to see what trends come about with changes in funding and changes in fundraising as indicators of individual donation levels. If suggestive trends are found, possible reasoning for those trends will be talked about. Using the previous literature on this topic, I will explore how fundraising reacts to government money, how individuals react to government money, and how individuals react to fundraising. Considering the addition of fundraising patterns in recent studies on crowding out, and the possible importance of it, the next chapter includes stati stics on such patterns as they are available.
Cut ler 33 Chapter 4 | Empirical Links On Addressing Public Goods While the public radio market cannot be modeled by a public goods analysis, there are some worthwhile suggestions made in attempting to examine public ra dio in the scheme of a public goods analysis, as government sources do help to provide for public radio's livelihood. By exploring aspects of public radio that can be addressed in this context, it is hoped that some socially optimal level of funding can be established. A public good in theory is traditionally under supplied due to free riders, and whe re public radio is non rival and non excludable in its output (and seems to model a pure public good), financial support for public radio comes from a mix of sources including government and private individuals whose motives are not evidently altruistic. Free riders in the public radio market consist of listeners who do not become members' of public stations. These listeners do not donate to the public statio ns that they listen to. More than half of public radio's listeners fall into the free rider category. The listeners of public radio who do contribute make up an average of around 35% of public radio's cash revenue. There are other sources such as grants from businesses, underwriting fees, and local and state government subsidies the federal government (through the Corporation for Public Broadcasting) makes up the difference. Some level of public radio is then provided for, socially optimal or not. There is some debate over the socially optimal level of government funding for public radio, discussed briefly in the Background section. Public goods theory explains that a public good would be undersupplied (especially in the realm of voluntary support) thoug h it has been argued by Kingma & McClelland (1995)
Cut ler 34 that public radio does not represent a wholly pu blic good. Consumer donations to public radio are characterized by Kingma & McClelland as being positively linked with the consumer's level of use of public radio services, which does not support a public goods hypothesis. Because a public radio consumer has some motivation to link their giving patterns with their use of public radio, the public radio good acts in some ways like a normal private good. Public radio consists of voluntary donations to a service by individual agents, but those agents act in a way that displays a choice of preference (they give to stations they like and want to support). The central findings of Walrus Research's Audience 98 study w ere that "listeners who rely upon public radio are more likely to give," and that "as reliance increases, so does the level of giving." (Bailey 2003) Later reports by Walrus Research helped solidify these claims. Reliance' on a public radio station was ge nerated by examining the results of The Public Radio Tracking study. The study re contacted individuals who were involved in Arbitron's market research study of radio listeners to ask them about their attitudes and behaviors with reference to listener supp ort and create a link between listening and giving. Reliance on public radio was determined by "purely behavioral variables like loyalty to a public station and becoming a core listener." (Bailey 2003) Loyalty refers to a calculation of an individual's lis tening to public radio as a percentage of that person's overall use of radio, where a core public radio listener is a person who uses a public radio station more than any other station. Uncovering these patterns is significant in establishing a positive re lationship between listening and giving, which disproves a public goods hypothesis that contributors are indifferent between a donation given by them and a donation given by someone else (ie. government entity) Private contributions to a public radio stat ion does not display an
Cut ler 35 altruistic action solely, but has been theorized to contain selfish and warm glow motivations. This relationship also helps to inform a more inclusive model of public radio giving patterns to determine if contributors to public radi o are indifferent between a dollar contributed by themselves or a separate source. Being truly indifferent would lead to conclusions in line with a public goods hypothesis that predicts a dollar for dollar crowd out of the private contributions of individu al agents resulting from an increase in government funding. It is empirically suggested that crowding out happens at a less significant rate (one dollar of government subsidy increase does not lead to one less dollar of private contributions) which is susp ected when considering the preferences that individual donors rely on when contributing. These factors provide support for impurely altruistic giving patterns. Public radio's output is non rival ; the cost of program production stays constant no matter how many people tune in to the program. The marginal cost to the producer will be zero. So, allowing for one extra consumer to access the station adds no additional costs. In this sense, the public radio good acts much like a pure public good. This specificat ion is important to note, because where some traditionally thought of public goods experience congestion problems at high volumes of consumers, it is not characteristic of public radio to do so. Further, Payne (2009) explains that government members who agree on the policies with which public radio is funded, are voted in by the same independent agents who contribute their private funds to the public stations of their choosing, thus creating a possible bias in the allocation of federal dollars. This infor mation helps inform public
Cut ler 36 radio funding studies, as adjusting for endogeneity errors between private contributions and government subsidies may be necessary. Changing Allocations, Changing Donations: The Study Private contributions to public radio stat ions have continued to float above 30% for over 10 years. In 2011, 34.8% of public radio revenue was from subscribers, up 11.3% from 2010's subscriber contributions which totaled 34.5%. Government contributions in the form of Federal Government Grants and Contracts and allocations from the Corporation for Public Broadcasting (these two entities make up the direct federal financial sources) moved from 11.8% of total revenue in 2010 to 10.8% in 2011 18 Looking at the aggregate figures above does not help to dr aw any substantive conclusions about the relationship between federal government funding sources and individual subscribers. It has been established that these subscribers are not indifferent between their personal donation and the government's subsidy, bu t these two are intrinsically linked in that those subscriber's tax dollars are dollars that partially fund public radio through the federal allocations (though it is a small percentage of tax revenues that are fed into the public broadcasting network). Co nnecting federal dollars to private giving could provide valuable information for further policy decisions on the importance of the continued allocation of federal dollars to a public broadcasting network. If voluntary consumer behavior changes with adj ust ments in federal dollars, the n any cuts or increases in federal allocations could have an impact on subscriber allocations. This is the link explored in this study. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! $& Public Broadcasting Revenue FY2011, CPB.
Cut ler 37 The Data Set Data for this study was procured from the Corporation for Public Broadcasting, as the source of financial figures on the revenues and expenses of public radio stations. The CPB data (in the form of Annual Financial Reports (AFR)) includes specific figures on individual public station's revenue sources, and includes a br eakdown of funding details. Yearly data on the amount of individual station's Community Service Grant allocation, total allocation from federal agencies, revenue from station subscribers, and subscriber development are all found in the AFR Schedule. The CP B's Annual Financial Reports also include data on the financials of station fundraising efforts, a variable suggested for examination in more recent studies on the crowding out effect by Andreoni and Payne (2011, 2009). Reports from the CPB were obtained f or a 10 year span (2002 2011) to track changes in funding amounts and can be used to account for time lags in the data. 2011 was chosen as an endpoint because 2011 is the latest available year of reported figur es. The AFR contains data on 400 unique public radio stations who are recipients of Community Service Grants from the Corporation for Public Broadcasting (though not all 400 of these s tations are represented by a 10 year span in financial figures, due to various reasons; changes in station structure, termination of CPB funds, and discontinuation of stations all account for fluxes in CSG start and end dates). Three hundred and thirteen of these stations report data for the continuous length of the 10 year data set, but data can still be used in the case where stations report less than 10 years of continuous figure s Because my analysis considers a year to year change (making for nine change variables), around 340 observations are able to be acquired from the data set per variable per year span. All monet ary variables are reported as real U.S. dollar amounts. Key variables from
Cut ler 38 this data set include CPB Community Service Grants (for radio unrestricted portion only), CPB Restricted portion of Radio Community Service Grants, Memberships and Subscriptions & Fund Raising and Membership Development. These variables come from two different AFR schedules AFR Schedule A and AFR Schedule E. Schedule A "is used to report all direct revenue recognized in the audited financial statements for the reporting year. Direct revenue includes gifts, grants, contributions of cash, payments for services, dividends on investments, royalty payments, etc." 19 Schedule E is used "to report all expenses and investments in capital assets." 20 Not included in this data set was inform ation on subscriber statistics indicating income or education levels, nor was station level data on this subject available. Defining the Variables i CPB Community Service Grants (for radio enter unrestricted portion only): This variable represents the gr oss revenue amount that a public radio station receives from the Corporation for Public Broadcasting, as part of the Community Service Grant program. The CSG program is how public stations receive public funding. The unrestricted portion of the CSG program is defined earlier, and makes up about 75% of CSG grantee funds. ii CPB Restricted portion of Radio Community Service Grants: This represents the gross CSG revenue as part of the unrestricted part of the program. About 25% of CSG grantee funds take this f orm. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! $# CPB, Financial Reporting Guidelines For Preparing The Annual Financial Report (AFR) & Financial Summary Report (FSR). Grantee Fiscal Year 2012 Edition. '% Ibid.
Cut ler 39 iii Memberships and Subscriptions (net of write offs): This is the revenue that a public radio station receives from private donors. This line is used "to report revenues from memberships and subscriptions that are, in aggregate, less than $1,000 per individual donor for the reporting year. Individual membersh ip contributions of $1,000 or more are considered major donor gifts and are to be reported," as Gifts and Bequests from Major Individual Donors The average amount of total donations (given in an amount less than $1,000 dollars per individual donation) ove r the years 2002 to 2011 was $ 2,510,738. GRAPH 4. 1 shows the average private donation made over the years 2002 to 2011. iv Fundraising and Membership Development: Fundraising consists of inducing others to contribute money, securities, time, materials, or facilities. [Some] activities whose costs, including salaries and benefits for personnel engaged in those activities," included in this classification are: Costs i ncurred in the solicitation of underwriting funds and grants; Costs of membership development; Acquiring and distributing fundraising material; Designing, printing, and distributing leaflets or posters for fundraising; Meetings for the purpose of improving fundraising techniques; Services of fundraising consultants and talent; Mailing costs related to fundraising; etc.
Cut ler 40 Figure 4.1 maps the average donation amount, a general trend, whose link to government contributions will be explored further in the next section. Figure 4. 2 shows the average CPB allocation across all stations. The aggregate measures provided in Figure 4. 1 and Figure 4. 2 show that both donations and allocations to public radio seem to be rising over the course the selected years. Addressing the validity of these trends forms the next component of this study. The AFR schedules are released yearly by the Corporation for Public Broadcasting, and detail each station 's self report ed financial statistics. Between the 0 20 40 60 80 100 120 140 Donation Amount (in $US) Figure 4.1: Subscriber Donation Avg All Stations 0 50000 100000 150000 200000 250000 300000 350000 400000 CPB Allocation (in $US) Figure 4.2: CPB Allocation Avg All Stations
Cut ler 41 years 2002 and 2005 (and possibly prior), The AFR Schedule A and its financial categories remained constant. The AFR Schedule A was modified in 2006 to include more categorical breakdowns, such as a new category account ing for the total amount of individual donors whose individual private gifts totaled over $ 1000 dollars. This breakdown stayed constant in the AFR Schedule A data sets between 2006 and 2011. So, there is some discontinuation in the data, as the 2002 2005 AFR Schedules and the 2006 2011 Schedules contain some variability in the brea kdown of revenue sources. O ne category in question Gifts and Bequests from Major Individual Donors was available between 2002 and 2011, though the sub category Total Num ber of Major Individual Donors was only recorded between 2006 and 2011. Because of this discontinuity, the variable was excluded from this study. An overwhelming majority of private funding statistics remained in the absence of this data. The data set allowed for the use of three main variables, Subscriber (or Private) revenue, Corporation for Public Broadcasting (or Public) allocations, and Fund Raising Expenses. According to the literature, public revenue level should have some effect on private reven ue accumulation, though a stronger correlation may be found between fund raising efforts and private donations. Again, not all 400 stations included in the data set included observations for each year, but with around 340 statistics observed for each per e ach year, the data remains telling. The CPB data was made most accessible by creating variables based on the change between years of figures, thus taking the ten year spans, and creating nine new variables calculated as: n (n 1) From Public Allocation, Pr ivate Donation, and Fundraising Expenditure, the variables Public Allocation Change, Private Donation Change, and Fundraising Expenditure Change were created. T his thesis topic is
Cut ler 42 concerned with the potential crowding out of private donation s such that Pri vate Donation Change is the dependent variable of this data set, and the other observations represent independent variables. My independent variables, Public Funding Allocations and Fundraising Expenses, are present for the 10 year span from 2002 through 2011. With this time span, I was able to create 9 variables representing the year to year change in public funding and in fundraising expenses (amount change between 2002 2003, 2003 20042010 2011). These variables offered a wide range of observations represented by figures over 1,000,000 dollars, and below negative 300,000 dollars. Figure 4.3 shows the frequencies of change amount in public funding for year span 02 03, broken into categorical variables based on that change amount. For reference, the frequency visualization for year span 2010 2011 has been included as well (Figure 4.4) There is some evident movement between funding levels 6 10 22 28 51 74 59 39 29 18 0 10 20 30 40 50 60 70 80 Figure 4.3: Frequency of Changes in Public Funding Between 2002 2003 Frequency Count
Cut ler 43 Eleven funding level categories were chosen to represent the continuous data of public funding changes, present in Table 4.5 Table 4.5 : Continuous Funding Levels As a more descriptive function, a means operation was performed, whereby the mean of the change in private contributions variable, SChange, was found for the categorical funding levels of my public allocation variable, AChange. For example, where category (1) saw six observed figures in 02/03 (those Public Contribution changes 18 35 37 52 67 35 23 24 25 33 0 10 20 30 40 50 60 70 80 Figure 4.4: Frequency of Changes in Public Funding Between 2010 2011 Frequency Count Category Change Amount 1 -100,000 or Less 2 -99,999 to -50,000 3 -49,999 to -25,000 4 -24,999 to -10,000 5 -9,999 to -1 6 No Change 7 1 to 9,999 8 10,000 to 24,999 9 25,000 to 49,999 10 50,000 to 99,999 11 100,000 or More
Cut ler 44 which fell within the range 100,000 or less ), the related figures in the private contributions variable (SChange ) were summated and divided by six to give us a mean of $152,315.67. The SChange means of all year spans for the 11 categories can be viewed in Figure 4.6 In most cases, some change between years was seen, and the categorical variable (6) representing No Change' was absent. To compare w ith the public funding changes, Figure 4.7 shows the private donations means accounting for the change in fundraising expenses (FRChange). The heightened averages at the poles of the graph and the generally smooth nature of the center regions show that a c hange in public allocation or a change in fundraising expenses is not accompanied by a hugely disproportionate cha nge in private donation levels. -300000 -200000 -100000 0 100000 200000 300000 Figure 4.6: Private Funding (Change) Means by the Change in Public Funding 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 2007-2008 2008-2009 2009-2010 2010-2011
Cut ler 45 In Figures 4.6 and 4.7 the most extreme donation means (SChange) happen in the case of extreme change am ount categories 1 and 11 of the independent variables. Table 4.8 shows the outer and inner public allocation variable categories. Table 4.8 : SChange Means Where AChange = 1, 5, 6, 11 -300000 -200000 -100000 0 100000 200000 300000 Figure 4.7: Private Funding (Change) Means by the Change in Fund Raising Expenditures 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 2007-2008 2008-2009 2009-2010 2010-2011 Year-Span -100,000 or Less -9,999 -to -1 1 to 9,999 100,000 or More 2002-2003 152316 32912 6409 122458 2003-2004 25283 23177 23307 79576 2004-2005 123512 -641 1392 10808 2005-2006 109984 -2204 2943 -14649 2006-2007 154927 12289 -396 97354 2007-2008 67090 14044 -1012 6180 2008-2009 100420 8767 20507 68339 2009-2010 -173216 29301 95366 207597 2010-2011 255997 16624 -5975 245470 Average 90701 14919 15838 91459
Cut ler 46 The similarity between the outer categories (with means 90,701 and 91,459) and inner categories (with means 14,919 and 15,838) show that there is some consistency between allocation amount from the CPB and member donations, though this could be some factor of station size (or other factors) and that financially large stations receive exponentially more revenue from both the CPB and private individuals. None of the averaged means at the bottom of the table are negative. As most of the year span means are pos itive, the table suggests that individuals donated similarly whether there was an increase or a decrease in CPB allocation amount. This may indicate that a pattern between government and individuals is non existent, while a possible pattern is emerging in regards to fundraising efforts Table 4.9 shows that a link between fundraising expenditures and private donations may be present in extreme cases of fundraising changes. Table 4.9 : SChange Means Where FR Change = 1, 5, 6, 11 Year-Span -100,000 or Less -9,999 to -1 No Change 1 to 9,999 100,000 or More 2002-2003 77219.09 14957.98 28963.35 34164 191472.21 2003-2004 -35126.42 19803.39 46858.09 13575.8 59808.42 2004-2005 17700.33 16318.17 34351.95 4508 266104.29 2005-2006 -265802 -13703.16 -35442.18 -19058 87352.3 2006-2007 -177293.39 6157.88 13271.25 5525.2 215528.54 2007-2008 67038.95 7750.12 614.875 -19712.25 147011.74 2008-2009 67270.48 -27559.78 2937.37 -61319.67 324165.17 2009-2010 106851.77 12954.52 11797.64 1675.25 236283.39 2010-2011 228325.07 29835.83 8974.15 12525.33 239428.58 Average 9576 7391 12481 -3124 196351
Cut ler 47 Where there was a significa nt drop of at least 100,000 dollars in fundraising expenditures, the SChange average over all years was 9,576, where an extreme increase of at least 100,000 dollars saw an average private donation change of 196,351. The link between fundraising and donatio ns is made most explicit by this data and by Figure 4.10 Figure 4.10 is a scatter plot of cross sectional data comprised of the by station average over all available years for the variables Private Donations and Fundraising Expenditures. The changes in private donations by fundraising levels seem to follow a generally posit ive trend (as shown in both the SChange means in Table 4.9 and the plot points in Figure 4.10 ). Some of the literature theorized that as non profits receive more public funds, less effort is spent on fundraising (and more of these funds go to directly supp ort organization expenses). (Andreoni and Payne 2008) Organizations with larger costs however, must fundraise according to their operating budgets, so looking at whether radio 0 1000000 2000000 3000000 4000000 5000000 6000000 0 2000000 4000000 6000000 8000000 10000000 12000000 Fundraising Expenditures Private Donations Figure 4.10: Cross-Sectional Average of Private and Fundraising
Cut ler 48 stations change their fundraising behavior based on allocation amounts (in a sta tic year comparison) should just prove that higher funding equals higher fundraising. Where Figure 4.10 shows that listeners donate more money where there are higher fundraising expenditure s (fundraising expenditures trended at a rate just below half of do nation revenue) Table 4.11 shows that fundraising, while not necessarily being a factor of CPB allocation amount, positively grows with CPB allocation amount. The equation: y 1 = 0 + 1 x 1 + e 1 where, y 1 = Fundraising Expenditures, 0 = a constant, c x 1 = Public Allocation Amount Expenditures), e = residual confirms that stations with a larger CPB allocation, also have larger fundraising expenses. TABLE X 9 includes no year to year change, but an average over all years of the two variables. The coefficient of 0.53 reveals that where Public Allocations (between stations) rise by one dollar, Fundraising Expenditures rise by around 50 cents. Again, I must stres s that this interaction may be rooted in various variables. Station size may be important, but considering that fundraising expenditures do not rise 1:1 with CPB allocation, the potential effects of fundraising crowd out should be considered in further stu dies and has previously been addressed by Andreoni and Payne in numerous studies. 21 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! 21 Payne 1998, Payne 2009, Andreoni and Payne 2003, Andreoni and Payne 2008, Andreoni and Payne 2011.
Cut ler 49 Table 4.11 : Fundraising dependent on CPB funding Dependent Variable: FUND_RAISING Method: Least Squares Included observations: 400 White heteroskedasticity consistent standard errors & covariance Variable Coefficient Std. Error t Statistic Prob. C 152113.1 32400.36 4.694796 0.0000 PUBLIC 0.532304 0.131218 4.056629 0.0001 R squared 0.428193 Mean dependent var 309134.0 Adjusted R squared 0.426757 S.D. dependent var 577147.2 S.E. of regression 436974.4 Akaike info criterion 28.81812 Sum squared resid 7.60E+13 Schwarz criterion 28.83808 Log likelihood 5761.625 Hannan Quinn criter. 28.82603 F statistic 298.0396 Durbin Watson stat 1.890555 Prob(F statistic) 0.000000 Testing for Crowding Out My research first suggested the creation of three variables, with observations taken from 400 public radio stations over the years 2002 to 2011. According to the crowding out hypothesis, changes in public funds will negatively affect private donations by some magnitude. According to recent literature, fundraising levels correlate more with private donations than public funds do with private donations, and fund raising and private donations should travel in the same direction. Instead of using the change variables I created for descriptive reasons, my regression function is an average of each station s three variables over the 10 year time span. New variables wer e created for the purpose of a cross sectional analysis. These did not consist of measuring the change
Cut ler 50 between years, but of finding the mean of one variable across all ten years for each of the 400 radio stations, x (2002 2011) A variable was created f rom these means, These measures were repeated for the appropriate observations. The three subsequent variables will now be known as Private Donations', Public Allocations', and Fundraising Expenditures.' Where my data allows for a limited scope in po ssible observations, private donations should be the result of a combination of public allocation amount and fundraising efforts (observed monetarily); the equation for private donations takes the form of: y 1 = 0 + 1 x 1 + 2 x 2 + e 1 w here y 1 = Memberships and Subscriptions ( or Private Donations), 0 = constant, c x 1 = Total CPB Community Service Grant (or Public Allocations), x 2 = Fundraising and Membership Development (or Fundraising Expenditures), e = residual. An Ordinary Least Squares regression will be used. Based on the literature and my descriptive statistics of the dataset, it is expected that 2 >0 or that the coefficient on the Fundraising Expenditures term should be positive. This is because a higher le vel of fundraising expenditures is expected to bring in a greater amount of private donations. The estimated Public Allocations coefficient will be left open, 1 >< 0, as this observation is testing the crowding out hypothesis. Observation of this variable will help to determine the interaction of government funding and individual public radio givers.
Cut ler 51 Where results are significant and the coefficient is positive, the data will suggest that there is no crowding out; crowding in may be present. Where the data is significant and the coefficient is near 0, presence of crowding out or crowding in will be limited. Where the data is significant and the coefficient is negatively correlated, crowding out may be present. I expect this variable to show a small, negativ e coefficient, which would be in line with Kingma (1989). Though, recent literature examines the connection between public funds and fundraising, and analysis of this estimation should also consider such interactions when evaluating the coefficients. Runni ng the regression, we find statistically significant results. Table 4.12 reports the regression results. For Public Allocations (PUBLIC), the probability is below the .05 significance level at p=0.0025 For Fundraising Expenditures (FUND_RAISING), the probability is below the .05 significance level at p=<.0000 The coefficient of Public Allocations is 0.27, and the coefficient of Fundraising Expenditures is 1.76. R 2 = 0.897, a very high result. With such significant results, careful interpretation of the results is necessary H eteroskedasticity measures were accounted for, and the regression remained consistent. This model however, may have been improperly conditioned. This regression, rather than inferri ng that changes in public allocation amounts may lead to changes in private donations, explains that larger stations receive a higher amount of funding from the Corporation for Public Broadcasting and they also receive a higher amount of private donations. Both of these measures can be accounted for by station size and station endowment', so to speak. It is my hypothesis that the independent and dependent variable results may be due to economies of scale implications. Where the traditional sense of econom ies of scale
Cut ler 52 applies to production costs at a high volume of good production, an economies of scale like analysis can also be applied to public radio markets. Bigger stations do have bigger costs well known anchors, reliable equipment, and higher quality segment production but a bigger station will also have an intrinsic value of quality that reaches higher than is possible for a small station struggling to keep up with local news. For this reason, the return on the good of radio' may be higher per lis tener for a more well endowed station. The use of the network of public radio programming may alleviate this effect, but all stations produce local content, and local content has the ability of strengthening or weakening a station in this way. Table 4.12 : Private dependent on Public and Fundraising Dependent Variable: PRIVATE Method: Least Squares Included observations: 400 White heteroskedasticity-consistent standard errors & covariance Variable Coefficient Std. Error t-Statistic Prob. C 38594.84 23467.22 1.644628 0.1008 PUBLIC 0.265229 0.087073 3.04605 0.0025 FUND_RAISING 1.763044 0.119952 14.69785 0.0000 R-squared 0.897109 Mean dependent var 661849.8 Adjusted R-squared 0.896591 S.D. dependent var 1213644 S.E. of regression 390275 Akaike info criterion 28.59456 Sum squared resid 6.05E+13 Schwarz criterion 28.6245 Log likelihood -5715.913 Hannan-Quinn criter. 28.60642 F-statistic 1730.732 Durbin-Watson stat 2.136013 Prob(F-statistic) 0.000000
Cut ler 53 Motivation to Change the Model In order to test against the above results (which I believe to be testing a size relationship rather than a dollar for dollar change relationship), a new model will be examined. Much like the variables created earlier, nine change variables were created fr om the original d ata set from each variable's 10 year span. These new variables mark the change in dollar amounts per year, per variable ( Private Donations, Public Allocations, & Fundraising Expenditures ). Twenty seven change variables (three variables per year change) with 400 unique station observations now comprise the data set for my regression. The model equation is the same as above, though the regression will now have to repeat nine separate times to accommodate for each year to y ear change. The hypothesis above remains such that the fundraising variable should be positively correlated with Private Donations According to the crowding out hypothesis, Public Allocations should be negative, though minimally correlated with Private Do nations Because this regression will look at the how the variables interact in the situation of a changing variable where CPB allocation increase s by $ 5000 for example we can see how private donations change along with it. By altering the variables to account for the changing interaction tracked across stations, rather than clumping the station means together, the results of this equation will more appropriately evaluate the possible presence of crowding out. The results of the 9 regressions are availa ble in Table 4.13
Cut ler 54 Table 4.13 : Correlation between Private Donations and the I ndependent V ariables 9 Year to Year Private Donations Changes Public Allocations Coefficient Fundraising Expenditures Coefficient DV Mean R Squared 2002 2003 0.111204 Prob =0.4939 0.182699 Prob=0.5143 42666.52 0.012855 2003 2004 0.204375 Prob=0.0276 0.072246 Prob=0.6877 32414.69 0.034772 2004 2005 0.041654 Prob=0.7574 0.451673 Prob=0.0320 38506.72 0.152245 2005 2006 0.047395 Prob=0.7963 0.326031 Prob=0.0215 16380.57 0.069707 2006 2007 0.564551 Prob=0.1896 0.697134 Prob=0.2028 15555.71 0.288368 2007 2008 0.403593 Prob=0.1850 0.615492 Prob=0.0347 22259.11 0.223923 2008 2009 0.077002 Prob=0.4927 0.524783 Prob=0.0189 22540.78 0.120044 2009 2010 0.696467 Prob=0.0001 1.139104 Prob=0.0504 27575.86 0.383112 2010 2011 0.169056 Prob=0.0453 0.097763 Prob=0.9015 64864.04 0.043228 Results There are multiple regressions to evaluate, with varying amounts of significances and coefficients. General trends to note are that Fundraising Expenditures were the most commonly significant probabilities <.05 (though one case was =.05). Including the .05 case, five Fundraising Ex penditure variables out of the nine total change regressions were significant. Three Public A llocation variables out of the nine total change regressions were significant. The significant public funding interactions were all positively correlated with private donations, with coefficients at 0.2, 0.7, and 0.17 Five out of nine of these
Cut ler 55 interactions were negatively correlated, thoug h not significant. The significant interactions are marked in bold. Also included in Table 4.13 are the dependent means, which track the average change in private giving over all years. There was a positive trend in private donations, except between 2005 and 2006. In that year there was an average decrease in donations by 16,380 dollars. The change means for the most part indicate positive growth of donations between each year span. Indicated by the means, some slump in individual private donations occurr ed beginning in 2005 and continued through (while growing slightly) to 2010. A large increase was seen between 2010 and 2011. The cause of this pattern is somewhat uncertain. General consumer confidence may have been rising in the years following the great recession, whereby listeners felt more comfortable contributing monetarily to a leisure like activity. The political environment during the time needs to be explored. Reason for the drop in donations between 2005 and 2006 (where steady growth was seen bet ween 2002 and 2005 ) is perplexing. Significant results suggest that there is crowding in of .17 .70 cents per CPB dollar. By checking the change function, rather than a 10 year average, we were able to glance into possible reactions to independent chang es opposed to general growth trends. The year span with the most significant results, 2009 2010, had a public coefficient of .67 and a fundraising coefficient of 1.14. The dependent mean was 27,576 dollars. Both of the coefficients were exceptional in rela tion to group, though the depen dent mean was hardly notable. I a m wary to conclude that crowding in is a definite occurrence based on my results and on limitations in my data; there are factors which were not controlled by the dataset such as income & educ ation of listeners, market share & reach of
Cut ler 56 programming, local vs. national content origins, etc. Variables like these should be held constant to empirically grasp the reality of a relationship between government and donors. Including these factors in one data set would be difficult due to the varying sources this information comes from. Still, in regards to my Public Allocations coefficients, the majority of negative coefficients surely question the validity of my significant results. Crowding out would s uggest a negative coefficient and the presence of a negative majority leaves any definitive conclusion open.
Cut ler 57 Chapter 5 | To Conclude The results in this paper support the implication that higher fundraising expenditures are related to higher levels of individual giving. This fact proves the need for steadfast fundraising efforts by public radio stations. A majority of the tested years showed that every extra dol lar spent on fundraising saw an average return less than 1 dollar (year 2009 2010 i s an exception at $1.13) but above zero These figures do not tell us whether more money was spent on fundraising than the gains from donations, but that u pon an increase in fundraising expenditures, there was a positive increase in donations received at a level below dollar for dollar Results regarding the emp irical validity of the crowding out effect were non supported, and potentially disproved. Donations tended to follow positively with the increased government dollars. This may be due to reputatio nal effects of stations those stations that show signs of quality may receive more money from listeners and government agencies in tandem. This study suggests that crowding in is more likely than crowding out. Where crowding out may suggest that donors are stubborn to maintain donations to an industry increasingly supported by their taxes, crowding in may imply that the public radio industry is reliant on public funds to support it, as donors may use government allocation as a potential signifier of rel iable organizations. Other possible causes for crowding in can be justified: i t may be that an extra jump in government dollars goes to support something such as new programming, or a revamped station website, that brings in a larger pool of potential donating listeners
Cut ler 58 In the second chapter, I addressed the 20 th century establishment of public radio, with its roots in the educational broadcasting system upon which radio was first founded. Emphasis was placed on how public radio is a necessary piece in the radio market as a subsidized entity in a market where educ ational broadcasting could not be sustained commercially. I summarized the public radio market today, explaining how public radio is funded, focusing on the dynamic between government and non government sources. It was important to note that the industry d oes not rely on one funding source in specific, though a large weight is put on the listeners, who make up around 1/3 of public radio's revenue. The breakdown of the total Corporation for Public Broadcasting funds is highlighted to explain the federal gove rnment's involvement in the public radio network. In the third chapter I formalize the basis of the theoretical work on crowding out. The literature review effectively begins by detailing Bruce Kingma's 1989 findings, which proposed that crowding out exis ts, and is around .13 cents per government dollar. This study was ground breaking due to the previously un accessed household level data that Kingma was able to link to station funding. Early work on the crowding out model was taken from public goods relat ed theories that individuals see government funds as substitutes for their personal gifts. I explain that the public radio market behaves in a more complementary model than suggested by a pure public goods hypothesis (which is purely substitutive), and a d ollar for dollar crowd out is almost impossible to prove empirically, though some degree of crowd out (such as Kingma's 13 cents) was not disproved. Further, non profit contributions were credited to exhibit Warm Glow' tendencies, and donors received some amount of utility simply in the act of donating, and thus, are not indifferent between their donation and the government's donation. This
Cut ler 59 behavior would lead to a crowd out less than 1 for 1. Kingma's study was a benchmark, and various studies came about testing his findings. Most studies are consistent in concluding that crowding in does not exist, but it's worth noting that different sub sectors of non profits (social human services, arts and culture, etc.) may react differently due to donor characterist ics. Here is where emphasis is made on breaking down the study of crowding out to using panel data sets or succinct studies of one non profit sector. Brooks (2001) explained the classification of direct and indirect crowding out, where direct crowding ou t looked at how changes in government allocations would affect donations, and where indirect crowding out may come from changes in the tax code. In arts and culture non profits such as public radio, an indirect crowding out may exist much more readily as d onors tend to have high income levels and high education levels, which are indicators of tax itemization. Brooks (2001) specifically addressed public radio, and concluded that public radio received higher contributions where taxes were higher as a way for contributors to forego some amount of taxes. The government indirectly contributes to this in foregone tax revenue. This link is not addressed in this study, though the consideration of this item will help to better model the theory. Finally, it was the work of Payne & Andreoni who concluded that crowding out was not totally attributable to government allocations, but rather, most crowding out was qualified by relaxed fundraising efforts. An organization recei ves the revenue it need s from a jump in govern ment dollars, and denies persistent fundraising efforts. Reasons for this may be that non profit watch groups see high fundraising costs characteristic of poor organizations, fundraising may affect growth of the actual service, and where revenue could be f ound in other sources, fundraising expenditure will try to be minimized. In
Cut ler 60 terms of radio broadcasts, the constant irking during pledge drive week may be something that stations would cut if they could. Andreoni & Payne (separately and together) find crow ding out to exist and that a percentage of it is attributable to the classic crowding out theory but that a high percentage of decreased donations come from decreased fundraising expenses upon realization of a government grant. The fourth chapter bui lt up the empirical study by explaining my data and methods. Contact was made with the Corporation for Public Broadcasting, and I was granted permission to use their Annual Financial Reports (Schedules A & E) to empirically asses s the existence of crowding out in the public radio market. I used information regarding individual station level figures on revenue sources and expenses to model the relationship of government funds to fundraising expenditures to private donations. I found statistically significant results showing that fundraising changes happen at a rate lower than CPB allocation changes, but the results indicate no form of crowding out. For every dollar increase from the CPB, there is a .53 cent increase in fundraising expenditures. However, as fu ndraising expenditures are defined in the CPB's self reporting guide currently and it is possible that 100% of the funds in this category are not specifically fundrais ing oriented, but more general in t heir inclusion of station development' expenditures. I originally modeled my study based on cross sectional variables created from the stations level averages over 10 years. A new variable was calculated from the mean of the 10 year span, and three inclusive average variables were used to see how private d onations reacted to different levels of government dollars and fundraising expenditures. These results which shows that private donations grew at a rate above 1:1 for fundraising,
Cut ler 61 and below 1:1 for CPB allocation, due to possible misspecifications in the m odel. This original model was a result of external controls such as economies of scale implications and listener dependent statistics (such as income level and education level, which are both indicators of higher donation levels), which were not included i n this data set. In an attempt to re calibrate the study, I created new variables which tracked the changes in all three variables, and performed nine new regressions accounting for each year change (based on my 10 original years). Where the results of t hese nine regressions were significant, similar coefficients were found in comparison to the original model. Where more than half of the public allocations coefficients returned as small negative values, none of the negative coefficients were marked by significant results. Based on significance alone, crowding in was found at a rate generally less than one dollar for both public allocations and fundraising. I believe these results are mis specified due to the earlier reasoning of data set limitations. Tracking the change amounts was too simplistic of a procedure, though controlling for station and individual variables is complex and strenuous. Appropriately modeling the crowding out hypothe sis would need to include, if not house hold level data matched to station level financial & market information (like in the Public Radio Tracking Study and in Audience 98), then possibly station level data consisting of financial statements and matched au dience analysis reports. My study included figures from every station that is financially supported by the Corporation for Public Broadcasting (the whole population of stations with federally allocated dollars). Averaging over the whole population meant that selectively looking for different effects at different station sizes was not possible
Cut ler 62 However, my study does find that the public radio market is growing. In the ten years over which observations were collected, eight of the year change variables sh owed positive private donation change means (out of nine ). The industry is growing and this aspect contributes to positive changes in both allocations and fundraising. Because I accounted for the change means, my results are not affected by the number of s tations included. In recent years, the existence of traditional crowding out has become questionable, though there is no recent evidence of crowding in either. The idea that public changes affect private giving has been reworked into saying that public ch anges affect fundraising efforts which in turn affect private donations. Where my study shows that fundraising efforts and private donations are positively linked, I do not have empirical evidence intrinsically linking an increase in government allocation to a decrease in fundraising expenditures. I found that fundraising expenditures did increase at a rate less than 1 for 1 with CPB allocation though this limited evidence cannot prove the dependence of these variables. I cannot prove that t hose extra CPB dollars are used to support station expenses where fundraising revenues (or other revenue sources for that matter; Grants, Local & State Government, Etc.) previously accounted for those expenses. A 1 for 1 change ratio would have potentially disproved such existence, and I conclude that the existence of fundraising crowd out is possible, though not explicitly supported by my results.
Cut ler 63 Bibliography 2012. "Corporation for Public Broadcasting FY2013 Radio Community Service Grant General Provisions and Eligibility Criteria." Corporation for Public Broadcasting. http://www.cpb.org/stations/grants/radio/generalprovisions/cpb_13RadioCSG_Gene ralProvisi ons.pdf 2011. "Public Broadcasting Revenue Fiscal Year 2010." Corporation for Public Broadcasting. http://www.cpb.org/stations/reports/revenue/2010PublicBroadcastingRevenue.pdf 2011. "Public Broadcasting Revenue Fiscal Year 2011." Corporation for Public Broadcasting. http://www.cpb.org/stations/reports/revenue/2011PublicBroadcastingRevenue.pdf 1940. "Sixth Annual Report, Federal Communications Commission." Washington, DC: United States Government Pringtin Office. http://transition.fcc.gov/ftp /Bureaus/Mass_Media/Databases/documents_collection/a nnual_reports/1940.pdf 1940. "Fifth Annual Report, Federal Communications Commission." Washington, DC: United States Government Printing Office. http://transition.fcc.gov/Bureaus/Mass_Media/Databas es/documents_collection/annu al_reports/1939.pdf 1946. "Eleventh Annual Report, Federal Communications Commission." Washington, DC: United States Government Printing Office. http://transition.fcc.gov/ftp/Bureaus/Mass_Media/Databases/documents_collectio n/a nnual_reports/1945.pdf Andreoni, James. 1990. "Impure Altruism and Donations to Public Goods: A Theory of Warm Glow Giving." The Economic Journal 100 (401): 464 477. Andreoni, James, and Payne, A Abigail. 2003. "Do Government Grants to Private Charities Crowd Out Giving or Fund raising"?" The American Economic Review 93 (3): 792 812. Andreoni, James, and Payne, A. Abigail. 2008. "Crowding Out Both Sides of the Philanthropy Market: Evidence from a Panel of Charities" (February). Andreoni, James, and Payne, A Abigail. 2011. "Is Crowding Out Due Entirely to Fundraising? Evidence from a Panel of Charities." Journal of Public Economics 95 (5 6) (June): 334 343. doi:10.1016/j.jpubeco.2010.11.011. http:// linkinghub.elsevier.com/retrieve/pii/S0047272710001714.
Cut ler 64 Arbitron 2012 Public Radio Today 20 11: How America Listens to Radio" Retrieved from http://www.prpd.org/Libraries/Digital_Media/publicradiotoday_2011.sflb.ashx Bailey, George. "Mega Trends: The Public Radio Tracking Study. Special Report." Greenbush, Wis.: Walrus Research, December 2003. B ailey, George. "The Public Radio Tracking Study Trend Report Winter Fall 1999." Greenbush, Wis.: Walrus Research, August 2000. Bergstrom, Theodore;, Lawrence; Blume, and Hal Varian. 1986. "On the Private Provision of Public Goods." Journal of Public Econ omics 2549. North Holland 29. Brooks, Arthur C. 2000. "Is There a Dark Side to Government Support for Nonprofits"?" Public Administration Review 60 (3): 211 218. Brooks, Arthur C. 2001. "Private Philanthropy and the Economics of Public Radio." SSRN Electro nic Journal doi:10.2139/ssrn.1808952. http://www.ssrn.com/abstract=1808952. Brunner, J. 1998. "Free Riders or Easy Riders"?: An Examination of the Voluntary Provision of Public Radio." Public Choice 97 (4): 587 604. Coase, R. H. 1966. "THE ECONOMICS OF BR OADCASTING." The American Economic Review 56 (1): 440 447. David Giovannoni, et al "Public Service, Public Support: Audience 98." Washington, DC: Corporation for Public Broadcasting, 1999. Estelle, James. 1983. "How Nonprofits Grow: A Model." Journal of Policy Analysis and Managment 2 (3): 350 365. Gazley, B., and J. L. Brudney. 2007. "The Purpose (and Perils) of Government Nonprofit Partnership." Nonprofit and Voluntary Sector Quarterly 36 (3) (September 1): 389 415. doi:10.1177/08997640062959 97. http://nvs.sagepub.com/cgi/doi/10.1177/0899764006295997. Hamilton, James. 2004. "Rationalizing Dissent Challenging Conditions of Low Power FM Radio." Critical Studies in Media Communication 21 (1): 44 63. Keeffe, Mary Ann. 1982. Public Broadcasting: Hi story, Organization, and Funding Washington, DC: Library of Congress, Congressional Research Service. Kingma, Bruce Robert. 1989. "An Accurate Measurement of the Crowd out Effect Income Effect and Price Effect for Charitable Contributions." Journal of Political Economy 97 (5): 1197 1207.
Cut ler 65 Kingma, Bruce Robert, and Robert McClelland. 1995. "Public Radio Stations Really, Really Not Public Goods: Charitable Contributions and Impure Altruism." Annuals of Public and Cooperative Economics 66 (1): 65 76. Manzo or, Sonia H., and John D. Straub. 2005. "The Robustness of Kingma's Crowd out Estimate: Evidence from New Data on Contributions to Public Radio." Public Choice 123 (3 4) (June): 463 476. doi:10.1007/s11127 005 7171 4. http://www.springerlink.com/index/10.1 007/s11127 005 7171 4. McCauley, Michael P. 2005. NPR: The Trials and Triumphs of National Public Radio New York: Columbia University Press. McChesney, Robert w. 2008. The Political Economy of Media: Enduring Issues, Emerging Dilemmas 1st ed. New York: M onthly Review Press. Milam, Lorenzo Wilson, and Thomas J. Thomas. 1988. The original sex and broadcasting: a handbook on starting a radio station for the community. San Diego Calif: MHO and MHO Works. Noronha, Frederick. 2003. "Community Radio: Too Many Ob stacles." Economic & Political Weekly 38 (17): 1628 1630. Payne, A. Abigail. 2009. "Does Government Funding Change Behavior"? An Empirical Analysis of Crowd Out" 23 (1): 159 184. Payne, A.Abigail. 1998. "Does the Government Crowd out Private Donations? N ew Evidence from a Sample of Non profit Firms." Journal of Public Economics 69 (3) (September): 323 345. doi:10.1016/S0047 2727(98)00005 X. http://linkinghub.elsevier.com/retrieve/pii/S004727279800005X. Stavitsky, a. G., R. K. Avery, and H. Vanhala. 2001. "From Class D to LPFM: The High Powered Politics of Low Power Radio." Journalism & Mass Communication Quarterly 78 (2) (June 1): 340 354. doi:10.1177/107769900107800209. http://jmq.sagepub.com/lookup/doi/10.1177/107769900107800209. Lise Vesterlund. 2006. Why Do People Give?" The Nonprofit Sector: A Research Handbook (2): 568 590 Vesterland, Lise. 2011. "Lecture 6: Empirical evidence on crowding out." Econ 2230: Public Economics. University of Pittsburgh. Pittsburgh, PA. Powerpoint Presentation. http://www.pitt.edu/~vester/2230_l6.pdf Witherspoon, John, Roselle Kovitz, Robert K. Avery, and Alan G. Stavitsky. 2000. A History of Public Broadcasting Ed. Steve Behrens. 2nd ed. Washington D.C.: Current Publishing Committee. http://www.current.org.