_moooo ||||| |||| || |||| || |||| || || ’;2||||| ||||||| || ||| This is to certify that the thesis entitled ?OTENT|A\. PENEMUE, (bf-"THE Top i—W'TY ’ltLEV lStbk) [\A ARKETS presented by #12 ants. C. l-ZATELL. has been accepted towards fulfillment | of the requirements for ' M 'A' degree in TM Wag. Mjpf Date W “a IZL/i/f’ 0-7 639 H 74 NOV 2. "“1 2002 POTENTIAL REVENUE IN THE TOP FIFTY TELEVISION MARKETS BY Denis C. Katell A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Telecommunication 1978 Approved<:;;;§L*/VE: ‘1X1342s Major Professpr ABSTRACT POTENTIAL REVENUE IN THE TOP FIFTY TELEVISION MARKETS BY Denis C. Katell This thesis seeks to initiate the use of potential revenue for the purposes of television market analysis. The term potential revenue is the product of a computed analytic constant multiplied by total retail sales of a television market (ADI). The study addresses itself to the purpose of pre- dicting potential revenue in each of the top fifty tele- vision markets, and to a lesser extent, economic viability of markets as it relates to entry of new television outlets. The potential revenue concept may be able to provide broad- casters, media specialists, brokers, and researchers with a measuring device which ascertains the financial and eco- nomic health of a particular tOp fifty market. A formula produced by this study could very well become an important tool in decision making for market television, including but not limited to: new entry, transfer of license, cable expansion, and operating expenditures. Denis C. Katell It is anticipated that this study will provide a base upon which an additional body of knowledge about tele- vision's market share of advertising revenue can be formulated. ACKNOWLEDGMENTS The author would like to express his appreciation to the following individuals for the parts which they played in the various stages of this thesis: John Abel, professor of telecommunication, Michigan State University. His patience with the author, his under- standing of the author's aims and goals, and his dedication to academic discipline made it possible for the author to acquire the understanding necessary for the preparation of the methodology employed in this study. Robert Yadon, the author's major professor. His interest, encouragement, and constructive criticism of the idea, its development and final fruition are appreciated and gratefully acknowledged. ii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . . LIST OF SYMBOLS . . . . . . . . . . . . . . . I. II. III. IV. INTRODUCTION . . . . . . . . . . . . . Purpose of the Study . . . . . . . . . Scope of the Study . . . . . . . . . . Terminology . . . . . . . . . . . . . REVIEW OF LITERATURE . . . . . . . Background of Television's Growth . . . Related Studies . . . . . . . . . . . . METHODOLOGY . . . . . . . . . . . . . . The Variables . . . . . . . . . . . . . Research Method . . . . . . . . . . . . Coefficients of Correlation . . . . . . Multiple Regression . . . . . . . . . . Discriminant Analysis . . . . . . . . RESULTS . . . . . . . . . . . . . . . . Introduction to Potential Revenue . . . Further Development of the K r Concept Discriminant Analysis of Entry . . . . Discussion of the Select Variables . . Discussion of the Discriminant Findings SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS summary 0 I O O I O O O O O O I O O O 0 Conclusions . . . . . . . . . . . . . Recommendations . . . . . . . . . . . A SELECTED BIBLIOGRAPHY . . . . . . . . . iii Page vi ll 12 16 16 27 38 38 48 48 48 50 53 53 62 66 67 72 76 76 78 80 82 Page APPENDICES Appendix A. The Top Fifty Television Markets in 1975 . . . 84 B. Individual Market Indexes . . . . . . . . . . 86 C. Potential Revenues and Efficiency Indexes for the Top Fifty Markets . . . . . . . . . . . 88 D. Analysis of Variance for the Upper and Lower 25 Markets in the Top 50 Markets Per Year . . . . . . . . . . . . . . . . . . . . 93 B. List of Markets Qualifying for New Entry by Year . . . . . . . . . . . . . . . . . . . . 94 iv LIST OF TABLES Table I Page I Growth of Television Households . . . . . . . . . 4 II Commercial Television Stations in Operation . . . 6 III Television Financial Data . . . . . . . . . . . . 19 IV Analysis of Variance for Entry in 103 Markets with Three or More Commercial Television Stations, 1966-1968 . . . . . . . . . . . . . . 31 V Correlation Matrix . . . . . . . . . . . . . . . 54 VI Values of "Kpr" Per Year . . . . . . . . . . . . 58 VII Analysis of Variance Between the Constant "Kpr" and Year . . . . . . . . . . . . . . . . . . . 59 VIII Regression Coefficients of Select Market Vari- ables Predicting the Dependent Total Television Revenue . . . . . . . . . . . . . . . . . . . . 63 Ix Prediction Results of Discriminant Analysis . . . 67 X Discriminant Coefficients and Amount of Rao's "V" For Selected Discriminant Variables . . . . 7l ADI IMI pr R +2 P R -2 TRSADI UHF VHF LIST OF SYMBOLS Area of dominant influence. The dependent variable Entry (see Equation (5)) Individual market index (see Equation (1)) Analytic constant (.0047) Variable defined by Equation (2) High potential revenue (see Equation (3)) Low potential revenue (see Equation (4)) Total television revenue (see Equation (1)) Total retail sales for the ADI Ultra High Frequency Very High Frequency vi CHAPTER I INTRODUCTION Because the number of television outlets in America have been increasing, and the available channels have been decreasing to a fewer number of Ultra-High Frequency (UHF) allocations in several markets, analyses of television markets have become a necessity for decision making. Whether it be a decision concerning construction, purchase, or capital improvements to existing properties, the gathering of necessary and relevant market data should be a prerequisite to any action. When commercial television stations were first licensed in the pre-war 19403 and then again in the post- war period of that decade, several economic problems pre- sented themselves. Major obstacles surrounded the absence of television receivers and a "built-in" audience ready to utilize the new medium. Other problems included, but were not limited to, inexperienced personnel, finding sources of programming, and developing a strategy for selling the medium to advertisers who committed a tremendous portion of their advertising budgets to print and radio media. It was the advent of television itself which had to create a demand for the medium in order for it to become as produc- tive and profitable as its predecessor, the radio broad- casting industry. As it turned out, television caused alarming social and economic waves across the country by the late forties. Restaurants and night clubs felt the impact. A variety series starring Sid Caesar and Imogene Coca, launched in 1949--1ater titled "Your Show of Shows"--became a Saturday terror to restaurants. It made pe0p1e rush home early. Television had briefly drawn people to taverns, but now home sets kept them home. Cities saw a drop in taxicab receipts. Jukebox receipts were down. Library circulations and book store sales were down. Radio listening in television cities were off; the Bob Hope ratings dropped from 23.8 in 1949 to 12.7 in 1951. The freeze kept sponsors on hand but the omens were frightening. The film world also felt the impact of television, and the 1948 Supreme Court decision to break up theater ownership of the large Hollywood studios added immensely to the economic upheavals relating to entertainment industries. Hosts of artists, assuming New York would be the capital for video production as it was for radio, headed eastward. Joining the migration were numerous fugitives from other media--newspapers, magazines, theaters, nightclubs, lecture halls.2 It seems as though they were all struggling for footholds in the new medium. It was a time for trial and error, success and failure. Amid a confusion of migrations and an atmosphere of upheaval, program experiments came and went.3 Although television outlets were new amdunpolished, the medium quickly took hold of the public, and the prestige of set ownership facilitated a tremendous volume of receiver sales.4 Yesterday, the business was more of an art. Today, it is more of a science relying heavily on audience, tech- nical, and economic research. The industry has become a highly refined entity, and several of the economic concerns have a new complexion. But for the most part, in the early days of television not many assumptions about the market had to be made. As long as the city slated for a new tele- vision outlet had a relatively large population and little or no competition from existing television facilities, viability was practically a certainty. Today, with the industry concretely established, and restrictions (both economic and regulatory) on new entry, the collection, processing, and analysis of market data are unavoidable for satisfactory managerial decision making. In light of rapidly advancing technology, the per- plexities of television economics continues its evolu- tionary path. When television began its growth, at the end of World War II, there were only 8,000 households with re- ceivers. Twenty-eight years later, television could be seen in over 68 million homes; 45 percent of those homes had two or more sets, and 70 percent of all homes had color sets.5 Since the modern era of television began in the middle of the past decade, several significant events appeared. A11 Very—High Frequency (VHF) allocations in the one hundred largest television markets have been taken, so new stations must operate in the UHF band.6 In the mid— 19705 the American Broadcasting Company, for years viewed as being a "weak sister, third network," has reached the number one rating position on a continuing basis. The network eco- nomic gap has been closed, and competition among the net- works is at its keenest point in history.7 The operation of independent television outlets have become more efficient as evidenced by recent annual financial reports by the Federal Communications Commission. Cable proliferation has become an increasing annoyance to broadcasters, and the advent of satellite distribution of video have caused new concerns. TABLE I GROWTH OF TELEVISION HOUSEHOLDS (in millions) Year TV Homes Year TV Homes 1949 1.6 1962 49.0 1950 5.9 1963 51.3 1951 12.4 1964 52.6 1952 17.3 1965 53.8 1953 23.4 1966 54.9 1954 28.2 1967 56.0 1955 32.3 1968 57.0 1956 36.7 1969 58.5 1957 40.3 1970 60.1 1958 43.0 1971 62.1 1959 44.5 1972 64.8 1960 45.2 1973 66.2 1961 46.9 1974 68.5 1975 69.6 Source: Broadcasting Yearbook (Washington, 1976), p. C-300. But the biggest economic gamble for television broadcasting continues to be construction of new stations in the larger markets. Adding to the present anxiety of several telecasters is the Office of Telecommunication Policy's VHF drop-in plan. The FCC has been requested to explore the possibility of permitting short spaced, lower powered VHF assignments in 70 markets.9 It is not clear yet as to the value of such facilities. The television industry is average sized.10 But, in the aggregate it looks lucrative to investors and potential owners. Even cliches such as, "a television license is a permit to print money," are commonly heard.ll However, in 1975 the network's 15 owned and operated television outlets alone captured 18.5 percent of the 669 commercial stations revenue.* Large market VHF stations tend to make handsome profits, yet 14 percent of all VHF and 49 percent of all UHF operations lost money that year.12 To the late comer, it is difficult to break into a highly competitive television market. Not only does the artificial scarcity of VHF assignments, cable penetration, and other technological aspects have impact on new entry viability, but also, and perhaps more importantly, economic room in the market to support an additional outlet must be a major concern. *Although 711 stations operated only 669 reported financial data. TABLE II COMMERCIAL TELEVISION STATIONS IN OPERATION Year Total VHF UHF 1945 6 6 - 1950 97 97 - 1955 439 294 117 1960 573 441 76 1965 586 487 99 1970 690 508 182 1971 696 511 185 1972 699 510 189 1973 700 511 189 1974 705 513 192 1975 711 513 198 Source: Broadcasting Yearbook (Washington, 1976), p. A-7. The proper evaluation of current market conditions is necessary for future television broadcasting growth, if not for the pure economics involved, then for the consider- ation of possible economic injury to existing stations caused by new entry. Economic success becomes increasingly dependent upon accurate evaluation of market conditions. This is particu- larly true with expansion into markets where competition with an existing station or stations could raise the ques- tion of economic injury. Precedent established for future Commission action on this matter is found in Carroll Broad- casting Company v. Federal Communications Commission: The Court of Appeals for the District of Columbia rejected the FCC's interpretation of the Supreme Court's 1940 decision in the Sanders Brothers case and made it mandatory for the Commission to con- sider economic injury protests when potential compe- tition seemed likely to affect the public interest adversely. In the related area of financial support for a new licensee, the Commission may have been too zealous in the hopes of providing maximum choice for the public. As the table on the previous page shows, there was a drop in the number of operating UHF stations between the years 1955 and 1960. The most significant cause for the decrease was, perhaps, the general lack of receivers equipped to tune the higher band of frequencies. As may be implied from the Communications Act of 1934 and the FCC's rules and regulations, there is a posi- tive burden of proof on every applicant to show that he has the financial resources to build and Operate a station. Al- though an ill chance of successful Operation of a new entry may be less critical than over burdening a market with extra competitors, the Commission does not disqualify applications in cities where there is obviously a lack of market char- acteristics to insure the probably success of the new entry. This, in part, explains why there are a number of stations which report losses year after year and still remain on the air. This may also be the most significant cause for out- lets to go "dark." In our free enterprise system, companies are lead by the profit motive. There may always be hope that the market characteristics will become more favorable and holding a license guarantees a place in the future market, but a disregard for market data compounds the finan- cial dilemma of struggling stations. The most advantageous time for an application to be filed with the Commission is at the point in time where the market appears that it can support additional service. However, this may not be the best time because there is a strong chance of competing applications in a market whose time has come. This project will attempt to provide an extra body of knowledge which can be used for the calculation of the potential revenue of a top fifty market, and in addition to other uses, it is hOped that determining when entry should be made can be reduced to a systematic process. Purpose of the Study This study will attempt to investigate relation- ships between various financial, economic, and market characteristics and the potential revenue for each of the tOp fifty television markets in America. It is anticipated that this investigation in addition to generating potential revenue data, will provide a reliable means of predicting whether or not new television service (most likely to be UHF) can be successfully sustained by the economic climate in each of the top fifty markets. But since the UHF handi- cap is eroding and the possibility of VHF drop-ins is increasing, this study will not discriminate against the type of transmission, but will focus on economic conditions and characteristics. In recent years, the analysis of total market revenue and the economic strength of radio markets have been enhanced by the use of a new term, potential revenue, a projection of the available revenue in a given market. To generate potential revenue as a variable, it is first necessary to select an index of relative market strength, and then to examine its relationship to existing 14 revenue data. In his thesis, Financial Behavior of Oklahoma Single Station Markets in 1973, Robert Yadon selected total retail sales as an indicator of market activity. The other criterion variable was revenue data per radio market published by the Federal Communications Commission. Yadon tested a random sample (N=30) of the top 300 radio markets and predicted that 95% of the nation's radio markets fell within two standard errors of the potential revenue index (Kpr") mean of .0039. A range of the potential was then formed between the two parameters; Kpr High = .0044 and Kpr Low = .0034.15 Potential revenue is the produce of a market's total retail sales multiplied by "Kpr," which is developed by dividing total market revenue by total retail sales. Reduced to simple terms, "Kpr" is the mean of percentages of total retail sales represented by the total market‘ revenue. Studying the ground work done by Yadon and its adaptation by broadcast investment experts including Paul 10 Kagan, have made a convincing argument that a study of potential revenue in television markets would be a worth- while undertaking.16 This thesis project will address itself to the pur- pose of attempting to predict potential revenue, and to a lesser extent, economic viability for new entry into each of the top fifty television markets in America. Although "Kpr" is a variable which will be a rudimentary calculation for the purposes of this study, the main concern will be distinguishing the best combination of variables to predict potential revenue for the various markets. "Kpr" in itself is somewhat powerful in providing a range in which a majority of markets are included. However, the range is not sufficiently narrow to prevent substantial discrepancies in the larger markets. It is the refinement of techniques for predicting potential revenue for tele- vision markets which this study hopes to accomplish. Although the "Kpr" and potential revenue concepts for radio are new and have not been commonly put to use in the industry, a study for television concerning their uses should be completed as soon as possible, so if similarities in both mediums are detected, simultaneous refinements in the scope and use of such techniques can be undertaken. Several related, but fundamental, measurements and analysis tools will be utilized, with the following objectives in mind: ll 1. To establish the relationship between total retail sales and existing total market revenue. 2. To establish the amount of potential television revenue in a market as a function of a specific variable or group of variables. 3. To project these findings as independent variables in order to indicate the probable economic viability of a market to the dependent introduction of a new television outlet. 4. On a comparative rank basis, how do the markets relate among themselves. 5. Does an increase in market size make it less likely for existing stations to capture the untapped po- tential revenue due to availabilities, time sales marketing, etc.? 6. To identify individual and groups of variables which merit more detailed analysis in the balance of television markets. Scope of the Study This study will be bound by some specific limita- tions which should be brought to light at this time: first, because of the investigative nature of the study, there should be no application of the findings contained in the research to additional markets without further testing in the balance of the television markets. The universe of the research shall remain the tOp fifty television markets. Second, this study does not attempt to define the "overall" success of any particular television market. The study will be primarily quantitative in nature, and in no way will account for all the qualitative aspects of some of the variables included. 12 Finally, being a preliminary investigation, the study will be descriptive in nature. That is, the study will describe the existing covariations, but will not, for the most part, attempt to test hypothetical relationships among the variables. Hopefully, this study will reveal quantitative indications of market performance compared to an established potential for the individual markets. "Kpr" certainly has the potential of becoming a welcomed accessory to the cur- rent devices used in financial decision making as it relates to market television. Terminology The following terms used in this study may require some additional classification: 1. Area of Dominant Influence (ADI). An exclusive geographic area consisting of all the counties in which the home market stations receive a pre- ponderance of total viewing hours. 2. Designated Market Area (DMA). Similar to the term ADI in that it is an exclusive geographic area in which the home market stations receive a prepon— derance of total viewing time. However, DMA is calculated by slightly differing criteria. 3. Efficiency Index (EFFIDX). A generated index which describes how close a market's actual revenue is to its potential revenue. 13 Individual Market Index (IMI). A calculated market index which gives, in percent form, the portion of total retail sales in the market which were spent on television advertising. Potential Revenue Index ("Kpr"). A generated constant created by the mean of IMIs in each year the study covers. Potential Revenue (Rp). A calculated amount of revenue which the market is capable of receiving relative to economic conditions and market characteristics. F OOTN OTE S lErik Barnouw, Tube of Plenty (New York: Oxford University Press, 1975), p. 114. 2 3 Ibid., p. 115. Ibid., p. 115. 4Leo Bogart, The Age of Television (New York: Ungar, 1956), PP. 90-91. 5Ward L. Quall and James A. Brown, Broadcast Management 2d ed. (New York: Hastings House, 1976), p. 135. 6Stanley M. Beson, The Value of Television Time and the Prospects for New Stations (Santa Monica: Rand, 1973), p. 1. 7Broadcasting (1978). "Special Report," January 2, 1978, pp. 28-29. 8Broadcast Management/Engineering (1976). "Satel- lites: Growth Competitor To Land Lines and Air Freight," October 1976, 12:10, p. 56. 9Broadcasting., op. cit. 10Federal Communications Commission (1075). "Broad- cast Financial Data--l975," Public Notice 68100 August 2, 1976. 11Quote attributed to Lord Thompson of Fleet. 12Federal Communications Commission (1075)., op. cit. 13Frank J. Kahn, Documents of American Broadcasting (New York: Appleton-Century-Crofts, 1968), p. 516. l4 15 14Robert E. Yadon, "Financial Behavior of Oklahoma Single Station Markets in 1973" (M.S. thesis, Oklahoma State University, 1975), p. 28. lSIbid., p. 39. l6Broadcast Investor (1976). (Rockville Centre, New York: Paul Kagan Associates, Inc., September 1976), p. 3. CHAPTER II A REVIEW OF LITERATURE Background of Television's Growth Television would be a more ordinary business if it were not for the fact that it makes use Of the electro- magnetic spectrum to transmit its signals.1 In 1927, Con- gress nationalized the spectrum, and ever since it has been regulated by Federal authority. License to Operate a broadcast facility is awarded and renewed by the Federal Communications Commission. The licensing power Of the FCC controls all activities of television stations. Even ad- vertising prices are indirectly regulated by virtue of the FCC's power to control entry and the number of competitors.2 Government policy toward television has shaped the performance Of the industry. The number of stations that are available to American communities, although greater than the number available anywhere else in the world, is smaller than the number that both advertisers and viewers would be willing to pay for.3 The scarcity Of television channels is due, basically, tO the fact that TV broadcast service must share a finite spectrum Of usable frequencies. Furthermore, no 16 17 two transmissions, whether Of the same or different kinds, can use the same frequency simultaneously in the same geo- graphic region. Consequently, one Of the most fundamental functions Of the Federal Communications Commission is tO assign the allocations of frequencies and bands Of fre- quencies to various types Of service using over the air transmission of signals. The amount Of spectrum space required by any type Of electronic transmission is determined by the amount of information which is carried. In the United States, each television channel is allocated a channel width Of six megahertz. In comparison to the standard broadcast band, one television channel represents nearly six times the amount of spectrum space used by the entire AM radio service. Therefore, the limitation Of providing a relatively small number Of Operating frequencies for television has caused the creation Of a market structure in which a small number Of firms supplies the major portion Of the indus- try's output. Since the resource Of spectrum space is relatively limited and because the government controls spectrum management among the several communication ser- vices, broadcasting has become an OligOpOly. The television broadcast industry is similar to other Oligopolistic industries in that in the absence of high barriers tO entry, high profits among existing firms should cause new firms to enter the industry. On the other hand, it is different from Oligopolistic manu- facturing industries in that under current regulation 18 by the FCC new firms cannot enter any market without a license from the Commission and existing firms typi- cally cannot expand their quantity Of output in response to rising demand. Economists agree that prices in Oligopolistic industries generally fluctuate less widely than those in more competitive industries. Each seller in the market will resist lowering prices because his few competitors will immediately match the cuts, leaving each with essen- tially the same share Of market and lower profits. Because Of time limitations (24 hours per day) and a regulated limit to the amount Of commercial time per hour, many more units of advertising may not be sold in relation to the new lower price. Because the price Of advertising is not per- mitted tO fluctuate freely, and since it cannot rest on the equilibrium point Of the supply and demand curves, a number Of advertisers will Opt to place their promotional expendi— tures in other media or not at all. Because Of this and the Federal Communication Commission's policy of not publishing individual station financial data, attempts tO analyze economic and financial behavior Of television broadcast stations or television mar- kets present unusual peculiarities not encountered in free (unregulated) enterprise systems. In competitive indus- tries, entry Of new firms take place until excess profits are reduced tO zero, and only sufficient profits are earned to market price Of capital and other inputs.5 In tele- vision, Commission decisions about spectrum allocation have l9 prevented entry from reducing profits in this way. Conse- quently, a television license is an extremely valuable asset, particularly in the larger markets. The television industry quickly became a viable medium as a carry over from the prior develOpment Of radio in the realm Of a commercial structure. By the end of 1950, the television industry as a whole became profitable. By the end Of 1953 television earnings equaled those of radio and have shown an almost unbroken advance thereafter. TABLE III TELEVISION FINANCIAL DATA (in millions) Year ' Revenues Expenses Income 1952 $ 324.2 $ 268.7 $ 55.5 1954 592.9 502.6 90.3 1956 896.9 707.3 189.6 1958 1,030.0 858.1 171.9 1960 1,268.6 1,024.5 244.1 1962 1,486.2 1,174.6 311.6 1964 1,793.3 1,377.7 415.6 1966 2,201.0 1,710.1 492.9 1968 2,520.9 2,026.1 494.8 1970 2,808.2 2,354.4 453.8 1972 3,179.4 2,627.3 552.2 1974 3,776.3 3,039.2 737.1 Source: FCC Annual Reports for each year listed. During the early 19503 radio earnings were being divided among an ever increasing number Of standard broad- cast stations, whereas television's growth, considering its potential, was much slower. The slow "closed market" 20 growth Of television enhanced earnings for those fortunate enough to hold permits to Operate stations in the late 19405 and early 19505. One Of the most significant causes in holding down competition and growth was the so-called "freeze" of tele- vision applications from 1948-1952.8 By the fall Of 1948, the Federal Communications Commission became increasingly aware Of the following conditions: First, the current allocation plan, adopted before much was known about VHF propagation, caused interference between stations; second, the twelve channels then allocated tO television would prove tO be entirely inadequate to take care Of the demand for stations. Furthermore, the color system question, which had clouded the issue all along, had become more and morg pressing as the technology of the medium progressed. The maximum number Of stations permitted tO Operate was arbitrarily fixed at 108. During the "freeze" the Commission stopped processing applications for television construction permits, and allowed only those stations al- ready authorized to broadcast or build a facility to con- tinue Operations. Throughout the "freeze" television receiver sales expanded at a rapid rate. The number Of sets rose from a quarter million to over 15 million.10 Holding the number Of outlets constant and increasing the number of potential viewers by enormous proportions has an extraordinarily large positive economic impact on the industry. The following summary clearly states what effect the "freeze" had on the development of television: 21 The patterns of ownership and control that developed among the first 108 pre-freeze television licenses set the trends for ownership in the ensuing post-freeze television proliferation. These lucrative stations were owned, for the most part, by radio licensees, including many Of the early radio pioneers, publishers, electro- nics firms, and motion picture interests. The high cost and high return of the pre-freeze television stations encouraged the rapid growth Of television group owners. The group owner developed out Of successful broad- casting ventures financing new broadcast investments; the profits from the first television station paid for more television stations. These patterns Of ownership, among the very profitable major market television pioneers, set during the freeze, began a process Of television ownership concentration that continued. The Federal Communications Commission held hearings, intermittently, until the later part Of 1951, to settle the engineering and policy questions that had brought on the "freeze." On April 14, 1952, the Commission issued the historic Sixth Report and Order.12 The new rules provided for 82 television channels. At the same time, the Com- mission issued a table Of channel assignments which made available more than 2000 TV channels in almost 1300 commu- nities throughout the United States, its territories, and possessions.13 In addition to the 12 VHF channels between 54 and 216 megahertz which were already in use, the FCC assigned 70 new UHF channels between 470 and 890 megahertz, Opening a new frequency band for the expansion Of the television industry. However, there are some inherent short comings in the transmission Of UHF signals. Probably the greatest limitation is the reduced coverage area Of UHF, requiring 22 additional antenna height and radiated power several times greater than VHF transmission to Obtain equal coverage. Another problem is the significant additional cost Of electrical energy to power high output UHF transmitters, which tend tO be considerably less efficient than trans- mitters Of lower frequency bands. Also, not tO be over- looked, is the purchase of UHF transmission equipment which can cost in excess of 100 percent Of the price for a VHF system. Little was known about UHF propagation characteri- stics in the early days. Additionally, little was known about the construction of UHF transmitters or high quality receivers of such high frequencies. However, since there was a need for UHF apparatus and a seemingly growing market for such equipment, manufacturers, especially RCA, were quick to produce acceptable products. But, UHF transmitters capable Of supplying enough power to create millions Of watts through high gain antennas were not available until the late 19505 and the early 19605. The refining and upgrading Of transmission equip- ment made and continues to make great strides in providing broadcasters more efficient, better quality transmitters. The weak link was and continues to be receiving equipment. The biggest drawback, in the fifties and most Of the sixties, was an absence Of receivers capable Of tuning the new UHF channels. TV set manufacturers, because of the low demand for UHF reception, continued tO build VHF only 23 receivers long after UHF was introduced. If viewers wanted to pick-up the higher channels, they had to buy separate tuning converters. It was not until 1964, when an act of Congress went into affect, that manufacturers were required to supply VHF and UHF tuning capability with every TV set manufactured in the United States or imported into this country for sale. Several organizations, comprised primarily of UHF broadcasters, have lobbyed the FCC for stronger regulations in the manufacturing standards Of television sets with respect to UHF reception. The Commission has taken steps to further erode the UHF handicap including the allocation of money tO develop a tuner with UHF/VHF parity. Other recent actions by the FCC including a relaxation Of UHF spacing will continue to erode the handicap. By the mid- 19805 UHF will probably be equal to VHF in tuning acceptabi- lity.14 Lately, Texas Instruments, a diversified elec- tronics company, has built a single UHF/VHF tuning unit which is currently being tested by the FCC.15 It is not unlikely that similar advances will occur in receiving antennas and lead in wire. Although the Sixth Report and Order provided much needed ground rules for the expansion of the industry, it contains a colossal misjudgement concerning ease of entry for new stations in most television markets. Dr. Allen B. DuMont, who headed a short lived fourth inter- connection service, the DuMont Television Network, made the 24 most cogent plea during the hearings process Of the "freeze."16 He argued that, insofar as possible, VHF and UHF channels should not be "intermixed"--that is, some areas should be assigned all VHF channels, while others should be assigned all UHF channels. DuMont wanted to eliminate the unnatural advantage enjoyed by the Older VHF frequencies and provide equality among all channels in a given area, permitting the growth of four networks with equal access to the public. DuMont's plan would have forced 12 pre-freeze VHF facilities to move to the upper band to accomplish non- intermixture.l7 These twelve outlets were pioneers who embraced television when others were belittling it, and they felt they should not be penalized by being banished to higher, and at that time unknown frequencies. Had the Commission adopted the DuMont plan, the economic and financial aspects Of television markets would compare strongly to that of the radio industry today, even though radio at one time had a group Of disadvantaged outlets in the Frequency Modulated (FM) band.18 Moreover, the DuMont plan would have more than likely, better facilitated the FCC's hope for increasing the number Of television stations in order tO expand the range Of choices available to the viewers in every locality. Although the remaining available allocations are UHF channels, and are presently less desirable than VHF frequencies to most potential investors, the Commission 25 has always hoped for an expansion Of service in local communities. This is the reason the Commission rejected one Of DuMont's alternative plans tO license a number Of super high powered VHF stations to cover vast expanses of the United States with one transmitter site in each Of a small number of geographic regions. For many years the Federal Communications Commission has attempted to increase the number Of television stations in the United States in order to expand the range Of choices available to viewers. Within the con- straints imposed by the requirement that television stations serve local communities, the Commission has sought in various ways, to improve the prospects for new stations. The policies pursued or considered have included limitations on the development Of cable tele- vision as a potential competitor to local broadcasting, the promotion of legislation to require that all tele- vision receivers be capable Of receiving UHF stations, and various attempts to change the frequency allocation plan to reduce the handicap that new stations would face. The struggle for "nonintermixture" of television markets has continued without much success since the "freeze" ended in 1952. On August 10, 1953, the FCC's table of TV allocations was upheld by a Court Of Appeals decision.20 In May Of 1954, a Senate subcommittee heard pleas from UHF Operators, who asked for a "hiatus" in the granting Of VHF permits, and other relief, including "deintermixture."21 However, the FCC defended the status quO and nO action was taken. During the following year, the Commission did consider the "deintermixture" Of Hart- ford, Peoria, Evansville, Indiana and Madison, Wisconsin, to make them all UHF only markets.22 Today, only the 26 Peoria television market Of the four listed above is an all UHF market. The latest quest for "deintermixture" was launched in late 1977.23 The licensee Of WCOV-TV, Montgomery, Alabama has petitioned the FCC to rewrite the "tortured history" of Alabama television. The plan would change the one commercial VHF allocation in Montgomery, the state capital, to UHF, and the adjacent market Of Columbus, Georgia would become an all VHF market capturing the reallo- cated Alabama VHF channel. The major contention Of the WCOV-TV plea is that the sole VHF station in the market has an unfair competitive advantage over the two commercial UHF outlets. According to the petition, the commercial VHF station accounts for 92 percent Of the market's total income for television. Judging the Commission's track record, it is unlikely action will proceed past the hearing stage. How- ever, it could possibly receive a fair review if WCOV-TV filed their complaint with the Federal Court. The common FCC practice is to recognize economic injury cases only in the awarding Of construction permits for markets showing substantial loss of income. Robert E. Lee, the most senior Of the Federal Com- munications Commissioners, initially appointed in 1953, has been a champion in the UHF development campaign. In the fall Of 1975, Commissioner Lee proposed the relaxation 27 of the multiple ownership and dquOly rules so that one licensee could Operate two channels in one market, if at least one channel is UHF.24 In his estimation, implemen- tation of such a rule relaxation would create more interest in the UHF band where there are 800 unused channels, just as FM was stimulated in the 19405 and 19505 by major AM stations. Today, FM radio has become very profitable and show signs Of becoming more popular than the AM radio service. Eventually, when UHF is recognized as an attractive investment, new interest will be encouraged and diverse ownership will be realized in the long run. Related Studies Up to this point in time the most extensive study which considers viability of markets and their viability Of supporting new entry is the Rand study Projecting the Growth Of Television Broadcasting: Implications for Spec- trum Use, 1976.25 Although the study was geared toward suggesting ways in which the Federal Communications Commis- sion could better manage and allocate UHF frequencies, the report Of necessity dealt with market by market economic considerations to predict UHF expansion and further non- television encroachment on assignments Of frequencies. The so-called "viable stations model" based on estimates Of relationships between the number Of UHF stations Operating in 1974 and the independent variables consisting of market 28 size, market number Of VHF stations, UHF penetration, wealth Of the market, and competition from out market stations was the major research considerations. The only basic financial consideration made was the use of a variable termed, market wealth (retail sales per household). There were three separate attempts to use individual station financial data without success. The Rand report prOposed to use profitability as an indicator Of economic viability. The first attempt was drawn from the FCC's work statement in its request for proposals, which suggested the estimation of television station revenues, partitioning these among market stations, and subtracting estimated expenses to arrive at profit pre- dictions. The second method involved direct estimates Of profits. The third, focused on station behavior, and con- tended that a station chooses its expenditure level to maximize profits. None of these methods did a very good job and were dropped from the study. But, even gOOd profit projections would have most likely been dubious indicators of viability, since there are a number Of stations which report losses year after year and still remain on the air. It is unfor- tunate that the financial methods failed. Had they pro- duced useful results, we would not be privy to how economic factors affect decisions to construct and Operate new tele- vision stations. 29 Station financial data would be a tremendous input to this project, if it were accurate. The major difficulty is that the financial data supplied to the FCC by stations are simple unreliable.26 General and administrative ex- penses are particularly susceptible to wide variations in accounting treatment. Since the FCC has no way to cross- check the statements against income tax returns, the seri— ousness of the problem cannot be assessed. Briefly the lesson learned from the Rand report concerning FCC financial reports is: The large variation Observed in the profits of appa- rently equally situated stations suggests that finan- cial data filed by individual stations have little usefulness for policymaking purposes. . . . comparisons of individual station performance are questionable because Of differences in station Operating modes and other factors that cannot be systematically taken into account.2 The lesser, but only true alternative method for explor- ation, would be extensive use Of market financial data. This method would be congruent with the widely accepted axiom--"When you buy (build) a station, you are buying (getting) a market." Douglas W. Webbink used analysis Of variance rather than regression tO predict new entry in his study entitled "Regulation, Profits and Entry in the Television Broad- 28 . . " AnalySis of variance was used because casting Industry. Webbink classified entry as a discrete variable which only takes on a limited number Of values over limited time such as +2, +1, 0, -1. 30 Webbink's model assumed that each firm including any new entrant Obtains an equal share of the market. He also assumed that all firms have equal operating costs, viewing audience, revenue, and profits. Analysis Of variance was performed on the means of five market variables including, net income, revenue divided by number of stations plus one, revenue over costs, total households divided by number Of stations plus one, and total households over number Of television stations. The independent variables consisted of four groups Of markets representing the number of entrants with the values mentioned above. Even though the model is naive in that it assumes equal share Of market for each station and equal costs, audience, revenue, and profits, it does fairly well in explaining the conditions necessary for entry. The main conclusion drawn by Webbink is that entry Of television stations is strongly affected by certain economic variables, particularly expected audience viewing size and therefore expected profits. He points out that the major markets where there are few if any allocations unused are the markets which can support and benefit from additional service. On the other hand, new entrants find it difficult at best to become viable in smaller markets where there are many Open channels. Therefore, it seems that the evidence shows that the FCC is mismanaging the allocation Of television channels. 31 TABLE IV ANALYSIS OF VARIANCE FOR ENTRY IN 103 MARKETS WITH THREE OR MORE COMMERCIAL TELEVISION STATIONS, 1966-68 . n = - l n = i 0 n = + l n = +2 F. variable (n=3) (n=76) (n=18) (n=6) y 37.88 148.9 759.7 1501.1 10.11** (72.61) (470.3) (1227.2) (944.4) R 1103 1609 3591 4685 7.67** 3:1 (543) (1544) (3743) (1934) 3 1.244 1.343 1.465 1.879 8.94** c (0.130) (0.258) (0.280) (0.176) T 67.99 123.9 225.6 233.2 5.83** 511 (24.91) (82.6) (198.7) (79.8) g 82.65 160.66 283.14 302.02 6.08** n (32.88) (103.70) (234.45) (104.13) Source: Regulation, Profits, and Entry in the Television Broadcasting Industry. ** Indicates that the means differ at a 0.01 level Of sig- nificance. The numbers in parentheses are standard devi- ations of the means. n = change in the number of commer- cial stations on the air from 1966 to 1968. n = number Of commercial stations on the air in 1966. 1966 market revenue of commercial stations in $ thousands. R C = 1966 market expenses Of commercial stations in $ thousands. T = 1966 television homes in thousands. y = R/n+1 - C/n Using published market data for markets in which there were three or more Operating television stations provided Stanley M. Besen and Paul J. Hanley in "Market Size, VHF 32 Allocations, and the Viability of Television Stations" with a regression model predicting the number Of television households needed to support a number of stations.29 The model started with the number Of television sta— tions in an all VHF market with unlimited allocations being a function Of television households. Since all mar- kets have only a limited number Of VHF allocations the path changes to a lower curve where the number Of homes in the market equals the number which exhausts the VHF allo- cations. James G. Saunders and Arthur R. Till conducted research in the mid-sixties which investigated the relationships between various station, market, and ownership characteri- stics and the financial behavior Of those stations. Although their report was focused on individual stations rather than markets as a whole, they did suggest some interesting possibilities for market exploration including: The correlation matrices resulting from this analysis should be used as raw data for studies using factor analysis and multiple regression. Detailed investi- gation should be continued tO determine the Optimum competitive situation in markets Of varying characteri- stics. From such studies it might be possible to define the maximum amount of broadcast service that might be expected by various kinds Of communities, and what kinds Of economic returns might be ansicipated by broadcasters providing these services.3 Much can be said for the usefulness Of market data. On one hand, it is the character of the market which will determine whether or not there is economic room for new entry. Market data is an aggregate Of individual station 33 data. The Rand study points out, " . . . comparisons Of individual stations are questionable but overall figures are useful." Stations and their individual audiences tend to fluctuate over time, but market size Of audience is relatively stable. In The Determinates Of Television Station Sales Prices, by Robert T. Blau et al., Indiana University, the findings showed that a station's sales price is primarily dependent on the level of annual net broadcast revenue.31 Saunders and Till found that corre- lations between audience size and the financial variables were high.32 We can conclude that it is primarily the recent ratings of a station that dictate recent revenue. This, of course, would be subject to fluctuation and change over time. The Rand report points out that the price Of an audience is due almost entirely to its size. They have found: That we can explain 75 percent Of the variance in "price" Of audience, strongly confirms the importance Of persistent market effects. In fact, it turns out that "price" of audience is sufficiently stable from year to year that one can do a pretty good job Of pre- dicting it by simply agguming that it is a constant in each market over time. Of course, it is not absolutely constant over time, but the above statement does indicate the stability and probably validity of market data. The relationship between viewership and the number Of local outlets is quite interesting and deserves attention in this project. Noll, Peck, and McGowan presented a regression that implied: 34 A single affiliate will attract between 42 and 45 percent of the potential viewers in its market. In a market with two stations, the total audience would be between 55 and 58 percent of the total potential, depending on the affiliate status of the stations. Finally, in a market with an affiliate of each network, the total audience is 60 percent Of potential.3 What, in fact, the finding is stating above is, that for each increase in service an unequal (diminishing) amount Of audience is added. This would seem tO be one Of the most important considerations to remember when contemplating the application for a construction permit to build a station. In summary, we conclude that market revenue is determined in part by audience size and that audience size does not change substantially from year to year. Since the audience size does not tend to fluctuate, changes in market revenue from year to year are controlled tO a larger extent by other market characteristics. Various handbooks published by the 0.8. Government including the Commerce Department, indicate that generally the most effective statistics tO use in market research are population, income, sales, and employment data. Out- lined in the next chapter are several independent variables which provide rather inclusive treatment of the above listed suggestions. As a preliminary tO hypothesis testing in scientific research, the author seeks to discover significant variables and how they relate to the balance of the variables, then lay the ground work for future testing Of hypotheses. FOOTNOTES 1Bruce M. Owen, Jack H. Beebe and Willard W. Manning Jr., Television Economics (Lexington, Massachu— setts: D.C. Heath and Company, 1974), p. 10. 2lbid. 3Roger G. Noll, Merton J. Peck and John J. McGowen, Economic Aspects of Television Regulation (Washington, D.C.: The Brookings Institution, 1973). p. vii. 4Douglas W. Webbink, "Regulation, Profits and Entry in the Television Broadcasting Industry," Journal Of Industrial Economics, XXII (1973)) p. 167. 5Bruce M. Owen et al., Television Economics, p. 11. 6Sidney W. Head, Broadcasting in America: A Survey of Television and Radio, 3rd ed. (Boston: Houghton Miffin Company, 1976), p. 209. 7Ibid. 8Ibid. 9Ibid. lOIbid. 11Lawrence W. Lichty and Malachi C. TOpping, American Broadcasting (New York: Hastings House, 1975), p. 146. 12Sixth Report and Order, Federal Communications Commission 17 Fed. Reg. 3905-4100, May 2, 1952. l31bid. l4Rolla Edward Park, Potential Impact Of Cable Growth on Television Broadcasting (Santa Monica: Rand Corp., 1970), p. 77. 35 36 15Broadcasting (1977). "Lee Sees TI Tuner As Spectrum Saver," December 12, 1977, p. 60. 16Johnathan C. Crawford, "Long Range Effects Of the Television Freeze," unpublished paper, Cincinnati, 1977, p. 4. 17David Lachenbruch, "The 3-Billion Dollar Gamble," TV Guide, November 1, 1975, p. 5. 18Charles Michaels, "A Study Of Selected Radio Markets," unpublished paper, Athens, Ohio, 1971, p. 9. 19Stanley M. Besen, The Value Of Television Time and the Prospects for New Stations (Santa Monica: Rand Corp., 1973), p. l. 20Broadcasting (1970). "A Play—by-Play Retro- spective," November 2, 1970, p. 114. 21 22 Ibid., 116. Ibid., 118. 23Broadcasting (1978). "Musical Chairs in Ala- bama," January 9, 1978, p. 37. 24Robert E. Lee, Television/Radio Age (1976). "The Drop-in Proposal: Let's Drop the Matter and Give UHF a Chance," March 15, 1976, p. 57. 25Rolla Edward Park, Leland L. Johnson and Barry Fishman, Projecting the Growth Of Television Broadcasting: Implications for Spectrum Use (Santa Monica: Rand Corp., 1976). 26Ibid., p. ix. 27Ibid., p. ix. 28Douglas W. Webbink, Op. cit. 29Stanley M. Besen and Paul J. Hanley, "Market Size, VHF Amocation, and the Viability Of Television Stations," Journal Of Industrial Economics XXIV (1975), pp. 41-54. \ 37 30James G. Saunders and Arthur R. Till, An Investi- gation Of Possible Correlations Of the Financial Behavior Of Broadcasting Stations (Athens, Ohio: Ohio University, 1966), p. 32. 31Robert T. Blau, Rolland C. Johnson and Kenneth J. Ksobiech, The Determinants Of Television Station Sales Prices: 1968-1973 (Bloomington: Indiana University, 1975). 32Saunders and Till, Op. cit. 33Rolla Edward Park, op. cit. 34Roger G. Noll, op. cit. CHAPTER III METHODOLOGY Since this study will be strictly ex-post-facto in nature, the data used will be taken from or generated from recorded economic, financial, and market statistics. Fred N. Kerlinger defines this type Of research as, "sys- tematic emperical inquiry in which the scientist does not have direct control Of the independent variables because their manifestations have already occurred."1 Data relating to the variables will be compiled from various published sources including the Federal Commu- nications Commission, U.S. Department Of Commerce, U.S. Department Of Labor, American Research Bureau, A.C. Nielsen Company, and the Standard Rate and Data Service, Inc. A relatively large number of variables will be employed based on the recommendations and findings of Saunders and Till in 2 1966. A few new variables, not included in the Saunders and Till study, will also be introduced. The Variables For the purpose Of this analysis, elements which are thought to influence total television revenue and the 38 39 potential revenue Of individual markets will be classified as economic, financial, or market characteristics. Vari- ables fitting the economic category will include total retail sales, consumer spendable income, buying power index, unemployment figures, and individual market index. Financial variables will include revenue, expense, income, and percentage Of spot revenue Of stations in each market plus per station revenue, per station income, poten- tial revenue, and an efficiency index. Market characteristics will include the variables important for market analysis but which do nO fall under the previous categories. These variables are households using television, persons viewing television, number Of television stations, number Of independent television stations, cable television penetration, and number Of tele- vision households. Except for the financial variables broadcasters have little if any affect on the last two classifications Of variables. The economic and market variables, however, dictate to a very large degree the values Of the financial variables. Although broadcast stations may improve revenue through management and sales techniques and income through management and efficiency techniques, the greatest impact will be expressed by economic and market characteristics. 40 The Financial Variables a. Total Television Revenue Total television revenue for each market is the total time sales of all television stations in the market less commissions to advertising agencies, representatives, and brokers, plus the sales Of programs, materials, facilities, and services, plus other broadcast income. This is perhaps the most important variable for consideration because it is the absolute gauge of television sales activity in a market. Collec- tively, it will represent the total gross return less commissions on investment and Operation Of television broadcasting stations in a market. Total television revenue will also be instru- mental in generating individual market indices. Divided by a market's total retail sales it will produce an index that will eventually be tallied across all markets in the top fifty each year to create a "Kpr" the mean Of individual market indices. "Kpr," the average percentage of total retail sales in terms Of total television revenue. A market efficiency index can also be gene- rated using total television revenue divided by potential revenue and will show on a comparative 41 basis how well a market does in trying to capture its potential in financial terms. Like the other published financial variables, total television revenue will be taken from annual financial reports available from the Federal Commu- nications Commission. b. Network Revenue Network revenue is the amount of compensation the stations receive from networks for transmitting programs from the network which include commercial messages sold by the network. Since the amounts of compensation received by stations from networks is for the most part determined systematically it should be-a good indicator of market strength. c. Spot Revenue Spot revenue consists of national and regional advertising dollars which buy time on individual stations. This revenue figure does not include commissions to agencies, representatives or brokers. Advertising agencies take into account a number of market characteristics when devising their national or regional advertising campaigns. It generally reflects a systematic method Of expendi- ture dictated by market size and other market characteristics. 42 d. Local Revenue Local revenue represents the amount of commer- cial time sold by stations to advertisers situated in the market itself. This figure also does not include commissions to agencies, representatives, and brokers. Local revenues are more elastic than network or spot revenues. Since local sales represent approximately 40 percent of total television reve- nue in 1975 nationwide, the amount of local revenue can indicate certain characteristics in markets including, but not limited to, competition from other media for local advertising dollars and sales management ability of local stations. e. Total Expenses These are the costs of doing business includ- ing technical, programming, selling, and general and administrative expenses. The amount of expenditure usually differs according to size of market, compe- tition, and commitments to public service. f. Total Operating Income Total operating income is the difference between total television revenue and total expenses and is figured before federal income tax. This in essence is the profit companies seek in a free enterprise system. Financially successful markets 43 would tend to have a relatively high total Operating income. A variable such as this could be useful in classification of markets as well as other stati- stical purposes. g. Potential Revenue Index A constant used to determine the potential revenue in each of the top fifty television markets. "Kpr" as it is abbreviated, is the mean of individ- ual market indexes and when multiplied with a mar- kets total retail sales produces the potential revenue for that particular market. h. Percentage of Spot Revenue Since there seems to be a systematic method in the placement of spot advertising dollars, the amount a market receives in respect to the total amount for the top fifty markets may reveal a partial explanation for the level of total tele- vision revenue in each market. i. Efficiency Index The product of total television revenue divided by potential revenue is an index of a mar- kets financial efficiency. The index can easily be changed to a percentage by moving the decimal point two places to the right, and shows how well a mar- ket did in meeting or exceeding the potential revenue of the market. 44 j. Per Station Revenue Although individual station data will not be used in this study, an estimate of revenue per station can be made to show the effects of the number of locally operating stations with respect to total television revenue. k. Per Station Income Again, individual station data will not be used, but a useful estimate of per station Operating income can be made to show the effects of the number of locally Operating stations with respect to total operating income in each market for each year. The Economic Variables a. Individual Market Index SMSA The individual market index (SMSA) is the percentage of total retail sales in the SMSA assumed by the variable total television revenue. As explained above, the index from all markets within a year create the constant value "Kpr," which will later be used to generate specific potential reve- nues for each of the individual markets. b. Individual Market Index ADI Same as above except that total retail sales are from the ADI rather than SMSA. 45 c. Buying Power Index Each year Sales Management Magazine list the relative buying power of SMSAs. By its very nature BPI is classified as an external market indicator. It is obtained by weighting the three factors of population, effective buying income (same as con- sumer spendable income), and retail sales and coverting them into a measurement of a markets ability to buy, and then expressing this as a per- cent of the nation's total buying power. d. Unemployment in Percent Assuming a market is plagued with high unem- ployment, total retail sales and other economic and financial variables could be adversely affected. The percentage figure rather than real numbers will be used since it is easier to manipulate an index than real numbers when the base (in this instance-- population) changes from market to market. e. Total Retail Sales SMSA This variable represents all sales and re- ceipts of all retail establishments. Retail esta- blishments are primarily engaged in selling mer- chandise for personal, household, or farm consumption. Yadon, as noted before, used total retail sales as an index of relative market strength with constructive results. Both Yadon's study and this 46 study extracted retail sales data from Standard Rate and Data Service's spot rate books. f. Total Retail Sales ADI Same as above except data includes SMSA plus the balance of counties comprising the ADI. Data for this variable comes from Spot Television published by the Standard Rate and Data Service, Incorporated. g. Consumer Spendable Income SMSA Consumer spendable income is similar to the Department Of Commerce term "disposable personal income," which is income remaining to persons after deductions of personal tax and non-tax payments to the federal, state and local governments. Data is for the SMSA and the source is Spot Television. h. Consumer Spendable Income ADI Same as above except data includes all counties in the ADI. The Market Characteristic Variables a. Households Using Television Households using television is the percentage of market households in the market with a television set turned on. In this study figures of HUT are for the Nielsen definition of market (DMA) which in most cases parallel the American Research Bureau definition. 47 b. Persons Viewing Television Persons viewing television is the actual estimated number of viewers watching. Both HUT and PVT will be taken from the July rating books of the A. C. Nielsen company for each market and year. July books are chosen to eliminate climatic dif- ferences in the various regions of the United States, since northern snow bound months tend to artificially raise viewing. c. Number of Television Stations Taken from the Broadcasters Yearbook, this variable not only accounts for the competition in a market but is also a relative indication of ability to support a number of television stations. d. Number of Independent Television Stations From the same source as the above variable, this variable is descriptive in terms of a market's ability to attract television service supplemental to the three network affiliates. e. Television Households The variable in which size and rank order of size is determined. It also indicates the potential audience of a market. Taken from ARB data. In this initial investigative study, there will be no attempt to include individual station characteristics. Perhaps at a later date, with the knowledge spawned by this 48 research, individual station characteristics may be helpful to investigate. For the purposes of this study the vari- ables listed above should prove to be fairly broad and relatively exhaustive. Research Method Each of the basic research objectives outlined in Chapter I will employ statistic tests which will be pro- cessed on the Michigan State University CDC 6500 computer system using various subprograms of the Statistical Package 3 for the Social Sciences created by Norman H. Nie et a1. Coefficients of Correlation To establish the relationship between total retail sales and existing total television revenue in each market, two statistical techniques will be employed. First, coeffi- cients of correlation will be computed to describe the degree and direction of the relationship. The Pearson product-moment coefficient of correlation (r) will be used for this purpose. Second, an analysis of variance will be used to determine the significance of the relationship from year to year across the six years 1970-1975, and the signi- ficance of the relationship considering the size of market. Multiple Regression For the task of establishing the amount of potential revenue in a market as a function of a specific variable or group of variables, multiple regression will be employed. 49 With multiple regression the researcher is allowed to study the collective and separate contributions of two or more independent variables to the variation of a dependent variable. As such, the two large purposes of multiple regression analysis are prediction and explanation, where prediction is really a special case of explanation.4 The stepwise method of multiple regression will be used because it yields the best prediction equation when all variables are entered into the equation. It also orders variables with respect to their contribution in explaining the variation of the dependent variable. Regression coeffi- cients reported in this study will be expressed in stan- dardized form. This will make it possible to specify per- cent changes in the dependent variable which can be attri- buted to percent changes in each of the independent variables. The output of the multiple regression subprogram to be used in this study is designed to supply regression beta weights (b) which serve as a means to identify the relative contribution of independent variables to a dependent variable. Once this reduction is complete, multiple regression analysis will be used in the prediction of potential market revenue, when only the select independent variables are known. This is accomplished by using the regression 50 coefficients to generate a formula that will estimate potential market revenue. Discriminant Analysis Discriminant Analysis will be used to indicate the probably economic viability of a market to the dependent introduction of a new television station. This type of analysis begins with the need to statistically distinguish between two or more groups of cases. In this study two groups, entry and non-entry will be considered for discri- minant analysis. The complete list of variables listed above will be used as discriminating variables in this subprogram. The mathematical objective of discriminant analysis is to weight and linearly combine the discriminating vari- ables in some fashion so that the groups are forced to be as distinct as possible.5 In other words, the objective is to discriminate the entry group from the non-entry group in the sense of being able to tell them apart. No single discriminant variable whether financial, economic, or market characteristic is expected to perfectly differentiate between the group of markets which had entry and the group which did not. But, by taking several vari- ables and mathematically combining them, it becomes rea- sonable to hope that there would be a single dimension on which entry markets are clustered at one end and non-entry markets at the other. 51 Since there is such a large number of variables in this study, perhaps more discriminating variables than are necessary to achieve satisfactory discrimination, the stepwise procedure will be adapted. Variables will be selected for entry into the analysis on the basis of their discriminating power using Rao's "V" method. "V" is a general distance measure and should provide the greatest overall separation of groups. The output of the subprogram will contain weighted coefficients which can be interpreted in the manner of factor analysis or multiple regression coefficients. Hopefully, it will be possible to determine the variables which best identify the markets, in this study, which are suitable for new entry. Comparing these markets with the 17 markets which in fact had entry should add validity to the use of potential revenue in determining when new entry should Occur and prove that a discriminant analysis is useful in establishing the criteria necessary for new entry. Once it is recognized that certain financial, economic, and market variables can, in a systematic way, did in the selection of markets which are ready for expanded service, the decision making process for investors seeking a market will be made more Objective in nature. FOOTNOTES 1Fred N. Kerlinger, Foundations of Behavioral Research 2nd ed. (New York: Holt-Rinehart-Winston, 1973), p. 379. 2James G. Saunders and Arthur R. Till, An Investi- gation of Possible Correlations of the Financial Behavior of Broadcasting Stations (Athens, Ohio: Ohio University, 1966). 3Norman H. Nie et al., Statistical Package for the Social Sciences 2nd ed. (New York: McGraw-Hill, 1975). 4Fred N. Kerlinger and Elazar J. Pedhazur, Multiple Regression in Behavioral Research (New York: Holt-Rinehart- Winston, 1974), p. 3. 5Norman H. Nie et al., op. cit., p. 23. 52 CHAPTER IV RESULTS Introduction to Potential Revenue Prior to the initial construction of a correlation matrix, additional variables need to be generated and tested. In order to weigh the economic strength of an existing market, parameters must be established. In this case, it became necessary to examine the economic viability of the market against some measure of potential revenue. To generate such a variable, it is first necessary to select an index of relative market strength, and then examine its relationship to existing revenue data. One well-known barometer of market activity is the variable total retail sales, as used by Yadon in his aforementioned study of radio markets. The other criterion variable, revenue data of the tOp fifty television markets from 1970 through 1975, is published by the Federal Communications Commission. Therefore it is possible to establish a rela- tionship between total retail sales (TRS) and the financial variable total television revenue (Rt) within each top fifty market across the six years of the study. 53 54 --- an. no. oa. av. oo. vo. aa. na.- av. av. av. an. on. oo. oo. av. an. av. na. oo. av. av. on. o.a ov. va. noon a an. --- ao. on. vn. oo.- oo. oo. va.- an. vn. vn. oo. oo. av. ao. an. no. on. vn. aa. no. on. no. an. on. na. and no. ao. --- na. oo. oo.- oo. ov. aa.) oo. no. on. an. oo. ov. oo. vo. an. oo. ea. oa. oo. oo. oo. oo. oo. nv. and ma. mo. na. -.. oo. oo. oo. oo. na.- oo. va. oo. aa. on. oa. va. aa. no. no. oo. on. na. aa. aa. oa. co. no. xoaanu av. vn. oo. oo. -u- vo.- aa. vo. oa.u oo. no. o.a on. on. on. vo. ao. oo. no. oo. oo. oo. oo. oo. vv. oo. on. on: non oo. oo.- oo.- oo. vo.- --- oo. oo.- oo.- oo.- ao.- oo.- no.) ao.- ao.- aa.- ao. oo.- oo. oo. ao.- oo.- vo.- oo.- oo. oo.- ao. max aa. on. ov. oo. vo. oo.- (.u ao. oo. aa. aa. aa. oo. oa. aa. ao. aa. vo. aa. no. oo. aa. aa. aa. vo. aa. vo. ocean aa. on. ov. oo. vo. no. ao. --- oo.- oo. av. no. on. aa. aa. oo. nv. ov. oo. oo. oo. nv. oo. av. aa. an. oo. ==>n na.- va.- aa.) na.- oa.- oo.- oo. oo.- a.. va.- na.- ma.) aa.- 6a.- aa.- ao.- na.- oo.- aa.o oa.- oo.- oa.- na.- ea.) 6a.- va.- oo.- nun mango av. an. oo. oo. oo. oo.- aa. om. va.- -.. no. oo. nn. an. an. oo. oo. oo. no. no. oa. ao. no. oo. vv. no. no. ao9 no o oo. ao. no. va. vo. aa.- ao. oo. oo.. oo. oo. vo. vo. oo. oo. -l- ao. no. ao. aa. ao.- oo. oo. oo. ao. no. ao. axnzo av. an. vo. aa. ao. ao. aa. mv. ma.) oo. oo. ao. on. on. an. ao. (u- ao. vo. aa. oo.- oo. ao. oo. av. ao. nv. an. an. no. an. no. oo. oo.- vo. ov. oo.- oo. no. no. vn. an. on. no. ao. -l- oo. oo. oo. oo. oo. no. on. vo. oo. n>a av. on. oo. no. no. oo. aa. oo. aa.- no. vo. no. on. vn. an. no. vo. oo. u)- no. oo. oo. oo. oo. av. oo. oo. no: na. vn. ca. oo. oo. oo. no. oo. oa.- oo. ma. oo. oa. on. oa. aa. aa. oo. no. t)- on. ma. oa. na. na. oo. oo. ao¢-aza oo. aa. oa. oo. oo. ao.- oo. oo. oo.- oa. ao. oo. oo.- ca. aa. ao.- oo.- oo. oo. on. .u- va. na. oa. oo. aa. oo. «wxm-aza av. no. oo. na. oo. oo.- aa. nv. oa.- ao. no. no. on. on. on. no. oo. oo. oo. ca. va. --- oo. oo. av. ao. ov. urooza nv. on. oo. aa. oo. vo.- aa. oo. na.- no. no. oo. oo. no. on. oo. ao. oo. oo. oa. aa. oo. .u- vo. nv. vo. am. an: a<99n on. no. oo. oa. oo. oo.- aa. av. ca.) oo. oo. oo. oo. oo. an. oo. oo. no. oo. na. oa. oo. vo. u)- av. oo. ov. 5mm aaooa o.a on. on. ca. vv. oo. vo. aa. oa.- vv. vv. vv. on. ov. on. ao. av. on. av. na. oo. nv. ov. av. .u- av. ma. sum noon ov. on. oo. oo. oo. oo.- aa. ao. va.- no. vo. no. an. on. oo. oo. ao. vo. oo. oo. aa. ao. vo. oo. av. -l- no. on: an: va. no. nv. oo. oo. ao. vo. oo. oo.- no. an. oo. on. an. an. ao. nv. oo. on. oo. oo. ov. no. ov. oa. no. .u-.xamwsnu.a NJD‘P 55 The top fifty markets were studied in light of internal properties inherent to the largest markets. Fac- tors such as number Of stations per market, and in turn the corresponding increase in competition, suggest that total revenue in these markets is nearly identical to the total potential revenue available. In select cases, total television revenue will exceed potential revenue at the expense of some other medium. Likewise, in some markets, the inverse would be true. However, considering the independent nature of each individual market and expecting a normal distribution, those markets exceeding the potential revenue available would tend to cancel those markets generating less revenue. The correlation coefficient between total retail sales for the ADI and total television revenue (R=.98), indicates a relationship that is positive, strong, and very dependable. In addition, the coefficient is significant at the .001 level, with over 95 percent of the common variance accounted for (r2 = .96). The correlation coeffi- cient between total retail sales for the SMSA and total television revenue was considerably less (r = 80) and was dropped from this analysis. The correlation coefficient for the top fifty tele- vision markets is slightly higher than what Yadon found in his sample of radio markets (N = 30) of 1972. This could be due to the fact that television time is bought more systematically by national advertisers and tends to be more 56 of a national advertising medium than local such as radio in most cases. This would tend to make television markets slightly more homogeneous in the respect that economic data and market characteristics influence to a large extent, the buying schedules of national advertisers and the amount of network compensation alloted by the commercial television networks. The other possibility of higher correlation might be attributed to this study's use of data over a six year period rather than for one single year as in Yadon's study. It is highly unprobably that the correlation coefficient was higher because of this study being a population study. Yadon's sample results were significant at the .001 level. A greater than chance relationship 999 time in 1,000. Appendix B enumerates the individual market index (IMI), or in other words, the percentage of total retail sales assumed by the variable total television revenue, for each of the top fifty television markets. Individual Market Index (IMI) = £§§ x 100 (1) t Given the nearly one-to-one linear relationship of the two criterion variables, total retail sales and total television revenue, it is possible to project these findings to any of the top fifty television markets with the gene- ration of a constant. The new variable, average potential revenue index ("Kpr"), is the mean of the individual market 57 indexes for all fifty markets each year. By using the new constant ("Kpr") in a top fifty market it is possible to generate a dollar-figure for average potential revenue (RP) available. Average Potential Revenue (RP) = TRS x "Kpr" (2) If the potential revenue index ("Kpr") is a measure of normality, then the potential revenue theory may be expanded to include the two outer parameters, high and low potential revenue expectation. Both the high and low potential revenue parameters equate to two standard deviations about the mean (Kpr)' When a sample survey is used, rather than a population study, the outer parameters equate to the standard error of the mean. At two standard errors (ZSEM), one can be 95 percent confident that all members in the population will fall between the two outer parameters. In a population study there is no standard error of the mean to consider because each and every case in the pOpulation (universe) has been tested. But, the standard deviation (SD), a measure of the variance about the mean is useful in ascer- taining the dispersion of cases. The interval of the two outer parameters is equal to two times the square root of the mean of the squared deviation scores about the mean of the distribution. 58 High Potential Revenue (Rp + 2) = TRS x (Kpr + ZSD) (3) Low Potential Revenue (Rp - 2) = TRS x (Kpr - 25D) (4) Only 12 cases of the total 300 (50 markets x 6 years) fell outside two standard deviations of the mean. In a sample, with the two outer parameter interval equal to two standard errors of the mean (referred to as the 95% confidence level) this would permit a total of 15 cases to fall outside the interval. It would be more advantageous to use a single "Kpr" for all years than a separate value for each year. Doing so would make future market studies using the "Kpr" concept more simple to manipulate and analyze, especially if trends across years is involved. To prove that it would be stati- stically acceptable to reduce the constants from each year into a single constant, it must be proven that the value of "Kpr" does not change significantly from year to year. TABLE VI VALUES OF "Kpr" PER YEAR 1970 - .004666 1972 - .004818 1974 - .004578 1971 - .004660 1973 - .004722 1975 - .004628 mean = .004679 standard deviation = .000083 As shown in Table VII, there is no significant difference in the value of "Kpr" from year to year. 59 TABLE VII ANALYSIS OF VARIANCE BETWEEN THE CONSTANT "Kpr" AND YEAR Sum of Mean . Source df Squares Squares F-ratio Between Groups 5 1.41.4 2.82-5 .103 p = .991 Within Groups 294 1.49’7 5.07"10 Total 299 1.49'7 Therefore, it is possible to use the mean of the six generated "Kpr" values in future market studies. To distinguish the new variable from its yearly values the quotation marks will be dropped. Thus, Kpr refers to the grand constant, which when rounded is equal to .0047. In other words, Kpr represents slightly less than one half of one percent of a market's total retail sales. It was also necessary to test the significance of "Kpr" across markets to insure that one calculated "Kpr" value per year did an acceptable job in establishing a reliable indication of potential revenue for each market in the tOp fifty. To facilitate the use of analysis of variance, markets one through 25 were placed in one group and markets 26 through 50 were placed in a second group for each year. For each of the years between 1971 and 1975, there was no significant difference in the values of "Kpr" for the first twenty-five and second twenty-five markets. Only 60 the year 1970 produced a significant difference in "Kpr" between the two groups. A summary table of the results of these tests can be found in Appendix D. In summary, Kpr (.0047) has been found acceptable to use in potential market revenue analysis for years 1971 through 1975. Although there was no significant difference in Kpr across years, the variance among the markets in 1970 suggest one "Kpr is not an acceptable average among mar- kets for that particular year. Perhaps two or three "Kpr"s should be calculated for the differing sizes of markets in the top fifty for that year. For the purposes of this study, it was decided to abide to the predetermined methodology plan and use separate "Kpr" figures instead of the grand constant (Kpr = .0047) with respect to multiple regression and discriminant analy- sis. Since this is a preliminary investigative study, the use of separate "Kpr" figures for each year should produce slightly more accurate results. As for the year 1970, only three markets fell out- side two standard deviations Of the mean (E = .0047, s = .00076). It was also decided to maintain the use of the 1970 "Kpr" for the sake of consistency of analysis across the years. The affects of a slightly larger variance in the variable for 1970 will be substantially reduced in analysis across years. Further investigation of the Variance in individual market indexes for 1970 produced interesting results. For 61 instance, when a mean was calculated for each groups of ten markets in descending order, markets 1 - 40 produced no significant difference. The fifth group of ten markets (41 - 50) had a mean significantly lower than the other four groups. Also to be considered is the fact that markets high on the list of the top fifty tend to remain there. The last ten or so markets on the list tend to change from year to year. If this study recognized any market which happened to occupy a position in the tOp fifty markets across the years instead of using the 1975 list of markets, results may have turned out to be more favorable. However, such a plan would omit data from the markets under con- sideration for at least part of the six years covered by this study. It is more useful to trace the performance of the present top fifty, even though markets such as Salt Lake City, Norfolk, and Wilkes Barre-Scranton changed positions on the list in a significant way, or did not appear at all in previous years. Even with the small disparity of 1970, using the six separate "Kpr" figures produced gratifying results. The variable potential revenue was generated by multiplying each market's total retail sales for the ADI with the "Kpr" calculated for the corresponding year. This produced a dollar figure of potential revenue for each market (see Appendix C). 62 Potential revenue correlated very highly with both total television revenue and total retail sales for the ADI. The coefficient of correlation between potential revenue and total television revenue equalled .98, showing a positive, strong, and very dependable relationship. Over 95 percent of the common variance was accounted for (r2 = .96). Be- cause of the way potential revenue is calculated it can be considered a function of total retail sales for the ADI. The correlation between the two variables equal .9998 and share .9996 of the variance. This finding very convincingly establishes the Kpr concept as an indicator of potential revenue. The rounding of financial data to the nearest thousand dollars probably accounts for the correlation not reaching the point of unity (1.0). A definitely strong and reliable relationship has been established between total retail sales of the individ- ual ADIs and total television revenue. It has also been confirmed that potential revenue is a function of total retail sales in each of the tOp fifty markets. Thus, it is possible to gauge the potential of each individual market in financial terms. Further Development of the Kpr Concept The stepwise method of multiple regression was used for the purpose of trying to produce more accurate estimates of potential revenue in the several markets. Regressions including all the listed and generated variables, and 63 regressions using only a select number of variables showing very little covariance were processed through the computer. Prior to running the first regression a scattergram (a graphic representation of two variables) plotting total television revenue against total retail sales was analyzed. The scattergram showed the top three markets to be skewed broadly and away from the balance of markets. It was then decided that in addition to a grand equation for all the markets, several equations for the various sizes of markets should be considered. The markets cases were segmented into five groups in light of how their plots grouped on the scattergram. None of the regression models added to the power of total retail sales to predict either total television revenue or potential revenue. All the equations were similar. Total retail sales entered first and showed an average coefficient of correlation higher than .99 with potential revenue and .97 with total television revenue. TABLE VIII REGRESSION COEFFICIENTS OF SELECT MARKET VARIABLES (ALL CASES INCLUDED) PREDICTING THE DEPENDENT TOTAL TELEVISION REVENUE Source df F-ratio b Values Total Retail Sales 1 5668.30 (p < .0001) .9907656 TV Households l 6.33 (p < .012) —.0307794 % of Spot Revenue 1 3.10 (p < .079) .01326610 Unemployment l 1.63 (p < .202) .02021246 Mean .29795875 64 Total retail sales was the only variable with an extremely high F-ratio. The other three variables were either slightly significant or not significant at all. In addition the other three variables shared very negligible covariance with the dependent variable. Since total retail sales for the ADI by itself accounts for a high correlation (r = .98) with total tele- vision revenue, and because over 95 percent Of the variance is accounted for, it was not too disappointing that the regressions did not add any predictive power to the concept. To increase the accuracy of potential revenue esti- mates it may be necessary to include per station data to the general analysis. The normal set of market characteristic variables could not do a very good job in supplementing total retail sales for the ADI. The correlation matrix at the beginning of this chapter shows a large amount of covariance among the variables one would assume to have the best possibilities in qualifying the market in terms of financial success and the ability to ascertain the possibi- lities of having entry occur. It is surprising to learn that the variable tele- vision households shares relatively little variance with other variables that would seem to be very dependent on the number of homes in the market. For instance, the number of television stations and the number of independent outlets as well as households using television and financial 65 variables correlated surprisingly low with television households. This is of special concern because the number of TV homes is the current way market size is measured. Perhaps television markets should be ranked by a variable or variables other than total television households. The remaining variables chosen for the regressions, percentage of spot revenue and unemployment, offered no significant support to the equations. Percentage of spot revenue correlated highly with the spot revenue variable, but correlated low with all of the other variables. For those reasons it was a likely candidate for inclusion in the regressions using the reduced set of variables. Unemployment seemed at first to be worthy Of inclu- sion, but added the least to the equations. Except in selected markets, unemployment does not have a meaningful impact on other economic variables in the market. This is probably due to various social programs which provide assistance to those out of work. Even in markets where there was a record of high unemployment, total retail sales, consumer spendable income, and market financial variables were not greatly affected. Total retail sales for the ADI, by itself, is a powerful indicator of potential revenue. This is probably due to the large amount of variance it shares with many of the other variables under study. It correlated very high with all the financial variables except for spot revenue 66 where there was a moderate covariance. Economic variables also correlated highly with total retail sales. Even the number of television stations and the number Of independent outlets correlate fairly high with the level of retail sales. The analysis of these relationships are, in them- selves, noteworthy. Discriminant Analysis of Entpy Discriminant Analysis was used to ascertain how effective several of the independent variables were in classifying the markets into those which had entry and those which had no new entry during the six years covered by the study. Instead of this subprograms usual applica- tion, that of predicting which group of several differing groups each unknown case would fall into, it was used to compare a market's ability to support new entry and the instances in which entry actually occurred. Since not much was known about the several indepen- dent variables and their predictive or discriminating ability, no prior stipulations were entered into the pro- gram. It was decided to allow the prior probabilities to be equal. That is, a case's probability of falling into one or the other group (entry or no entry) was equal. For maximum separation in the distinction of groups Rao's V method was used. This provided a general distance measure based on the largest changes in "V." 67 The results were very encouraging. Although the analysis output advised there were five instances where entry should not have occurred, understanding some of the peculiarities of entry can explain the output in a manner which could not be considered by the computer program. Table IX shows the number and percentage of actual and predicted group membership. TABLE IX PREDICTION RESULTS OF DISCRIMINANT ANALYSIS Actual Group Predicted Group Membership Membership Name Cases Group 1 Group 2 Group 1 NO Entry 283 231 52 81.6% 18.4% Group 2 Entry 17 5 12 29.4% 70.6% Group 1 centroid = .08841 Group 2 centroid = 1.27748 81.0 percent of known cases correctly classified Chi-square = 115.320 p < .001 Discussion of the Select Variables The variables which provided the most discrimination were (1) number of independent outlets, (2) potential revenue, (3) total expenses, (4) local revenue, and (5) total retail sales for the ADI. The top discriminators include variables from each of the three variable categories outlined in chapter three. 68 The market characteristic, number of independent outlets, is the best discriminator variable. The markets which in reality have the most entry are the larger markets which already tend to have one or more independent tele- vision stations. During the six year period 13 new stations were constructed in the tOp 25 markets and only four were built in the second 25 markets. Of the 13 markets in the top 25 only Kansas City and Sacramento showed no previous entry activity. The other four markets which were in the second 25 showed no prior entry activity. Potential revenue was found to be the second best discriminator. But, since the term is expressed in dollars which is dependent on the size of the market and not solely on the ability to permit entry, a better way of analyzing potential revenue is in terms of an index. The efficiency index is an indicator of how far above or below a market is to the potential revenue esti- mated for that market. The index mean for all cases is .9986, meaning on the average markets are slightly below their potential. The index mean for non-entry markets is .9947, a modest decline from the mean for all cases. On the other hand, the index mean for markets with entry is 1.0636, an appreciable increase over the average. Therefore, markets which tend to exceed their potential are more likely to have entry than markets which fall below their potential. This is an indication that an entrepreneur would rather enter a market with excess 69 television advertising dollars and try to cut himself in for a share of the revenue currently flowing into the market, rather than enter a market which has untapped revenue. The proven success of a market appears to have more appeal to those who construct new stations than markets which seemingly have uncommitted advertising dollars. If this is true, and new stations are not constructed in markets with low efficiency indexes, there would be few, if any, actual observations available to see if a new entrant can actually capture the supposed untapped revenues. Total expenses, is more difficult to explain than the previously mentioned discriminators. Although all expenses of operating television outlets in a market are not fixed, in a broad sense, they can be thought of as approaching the behavior of fixed expenses. The costs Of doing business, discounting inflation, do not change sub- stantially from year to year. Costs in larger markets tend to be higher, but the revenues are also higher. However, there tends to be a leveling off of expenses, and a situation of increasing returns takes place. This is most likely why per station income seems to systematically increase as the markets become larger. Income data pertaining to a market selected for analysis prior to considerations for new entry is always studied closely. Markets below their potential most probably show 70 income levels below their potential, and in turn, discourage development of new outlets. Not only does this account for new entry occurring in high efficiency index markets, but also in part explains why the larger markets are more favorable for new stations. Local revenue and total retail sales for the ADI complete the list of the best discriminating variables. Both variables are, to a great extent, prOportional to the size of the market they represent in each case. Since it has already been established that larger markets are most likely to experience new entry through analysis of the aforementioned variables, it is not surprising that these two variables rated as high as they did. Since all new entrants in the top fifty markets could not obtain affiliation with an established television network, it would of necessity make them independent out- lets. Since network compensation would be omitted from the categories which comprise total revenue, local revenue becomes increasingly important to the entrant. Histori- cally, independent operations have depended heavily on local advertising revenue to sustain their Operation. It has been only recently, with the sharply rising network rates and spot costs of large affiliated stations that national advertisers have seriously looked at the field of independents for increased placement of spot advertising. Even with the dispersion of spot dollars it 71 will be quite some time before independents can rely less upon the local advertising commerce. So, it would seem likely that potential entrants would look very closely at the amount and quality Of local advertising in the perspective market. Although television is generally viewed as a national advertising medium, inde- pendent stations are more like the radio broadcasting industry, in that local sales are extremely important for survival. Total retail sales for the ADI is a direct factor in calculating each market's individual market index. It also, in a more indirect way, is heavily responsible for each market's efficiency index. As a gauge of relative market strength, retail sales becomes an important con- sideration to investigate before plans to enter are made. TABLE X DISCRIMINANT COEFFICIENTS AND AMOUNT OF RAO'S V FOR SELECTED DISCRIMINANT VARIABLES Variable Coefficient Rao's V Significance N of Ind. Sta. -l.l46960 11.7739 .001 Potential Rev. .293097 6.1909 .013 Total Expenses -.022648 3.2534 .071 Local Revenue .099270 2.6654 .103 Total Retail Sales -.001011 2.4735 .116 Constant -1.572800 Group 1 centroid = .08841 Group 2 centroid = -l.27748 72 Prediction of the dependent variable, entry, is not possible through utilization of the unstandardized discriminant function coefficients for the three independent variables, and the constant, in a weighted equation. Con- sider the possibility of expressing entry (E) as a function of the independent variables where E = -l.5728 + (number of independent station x - (5) 1.14696) + (potential revenue x .293097) + (total expenses x - .022648) + (local revenue x .09927) + (total retail sales x - .001011) Application of this formula for any market in the top fifty television markets should provide a reasonable estimate of market readiness for entry. The closer the dependent variable (E) approaches -l.27748, the centroid for the entry group, the closer a market comes to the criteria needed for entry. It should be stated again, that this formula con- siders the economic room in a market for new entry. In no way can the use of this formula insure successful financial operation of a new television broadcasting station. Discussion of the Discriminant Findings In this analysis 52 market cases were categorized as being eligible for entry. Of this group 12 markets actually did have entry. But, five markets had entry when, according to the discriminant analysis, they were not sta- tistically eligible. 73 There are several reasons why actual entry was not nearly as high as the predicted entry. First, since the study covers six years, there were several instances when a market was repeatedly eligible for entry year after year. For example, San Francisco was sited in four of the six years for new entry. If a station went on the air early in the six year period, the change in the variable number of independent television stations may have kept the market from reappearing in subsequent years. If a market had no new entry the probability for entry in subsequent years would be higher with all other variables held constant. Second, some markets such as San Francisco, have no TV channel allocations which are unused, or in the case of other markets the only allocations remaining are very high on the UHF band. Finally, certain qualitative reasons such as competing applications could account for markets having no entry in the year or years new entry was predicted. In the market cases which shouldn't have had entry but did, rational explanations can explain why they should not be considered as market cases in default. The markets Baltimore and Indianapolis both had new entry in 1971. In that year both markets were predicted to have no new entry. However, Baltimore and Indianapolis were selected by the computer for entry the year before. Considering the lag in applying for a broadcasting license and the problems 74 inherent in building a station, we can assume the default was caused by unforeseen time delays. The Grand Rapids/Kalamazoo market was never slated for entry by the analysis. However, in 1971 a new entry occurred in Battle Creek, a city of considerable size 5 adjacent to the other two cities. In the case of markets which have stations in more than one city, special con- siderations must be made for new entry. The reasons would appear to be more qualitative than quantitative. Since allocations for television channels are on a city not mar- ket bases, expansion is likely to occur in instances where there is more than one major city in the market. In this particular case, the new entrant was able to Obtain a net- work affiliate contract from ABC, although there already is an ABC affiliate in the market. Houston and Orlando had new entry in 1971 and 1974 respectively. Houston was slated for new entry twice, in 1974 and 1975. Orlando was never predicted to have entry. Both markets however, are in tremendous growth areas. Since it is an acceptable practice to anticipate the time when a market will be ready to support a new television station, it is not uncommon for an enterprise to construct an outlet early to avoid competing applications for a specific channel. Such could very well be the case for both Houston and Orlando. The two growth markets undoubtedly will be able to support the entrants at some future time. 75 Another aspect to this consideration is that the early entry of a station will most likely cause a delay in the time when another entry can be warranted, and thus, ward off potential future competition. With a bit of rational explanation, the markets with unpredicted entry can be accounted for satisfactorily. It should be stressed that the explanations above do not completely justify entry in those markets. They were offered as post analysis considerations. If it was fairly uncomplicated to enter specifications for several qualita- tive considerations into the computer program, the number of unpredicted entry cases would be reduced. An overall efficiency index analysis shows that larger markets in the top fifty are more likely to meet or exceed potential revenue projections. Furthermore, markets at the lower end of the list seem to do fairly well with the exceptions of a few markets such as Providence, Memphis, Greenville-Spartanburg-Ashville, Grand Rapids-Kalamazoo- Battle Creek, Charleston-Huntington, Harrisburg-York- Lancaster-Lebanon, Wilkes Barre-Scranton, and Norfolk- Portsmouth-NeWport News-Hampton. It appears that hyphenated markets are the ones most likely to experience a hardship in capturing untapped revenues. CHAPTER V SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS Summary The purpose of this exploratory study was to help determine which independent variables were the best pos- sible indices of potential revenue in the tOp fifty tele- vision markets. Of necessity the relationship existing between total retail sales and total revenue in the radio study done by Yadon was adopted for use in this television study. Another purpose of this investigation was to see how accurately a discriminant analysis of market variables could predict markets eligible for new entry. In addition, it was possible to develop a formula, through discriminant techniques, to predict the economic room available in markets as it relates to new entry when certain select financial, economic, and market characteri- stics are known. The problems of determining potential revenue and ascertaining which markets were eligible for new entry relied solely on the use of market data. Specific station data cannot be obtained from the Federal Communications Commission. An extremely high voluntary compliance rate 76 77 from individual stations to the authors request for finan- cial data was not expected, so there was no attempt to include such data in the study. This is probably due to various reasons which are not necessarily shared among the specified markets. In Greenville-Spartanburg-Ashville and Harrisburg-York et a1. there are five network affiliated stations. With only three commercial television networks available, duplication in affiliated outlets occurs. The reason for duplication may be due to wide geo- graphic separation of cities within the market. In addition the combining of several SMSAs into a single market produces a hetrogeneous ADI in respect to retail trade and other economic aspects. It is usually the case, that one or two VHF stations cover the entire market. The outlets that do not cover the entire area, are usually located at the fringe of the market area. Thus, the disadvantaged stations cannot charge advertising rates commensurate with the ADI total retail sales, population, television households, or other economic indicators. 50 there are basically two factors which are easily recognizable in hyphenated markets. The first is wide geographic dispersement of cities within the ADI, the second concerns the age old problem of intermixture of VHF and UHF services. Since the markets which had low efficiency indexes were not the larger markets, it is almost impossible to 78 determine if big markets need new entry to capture untapped revenue. The only tOp ten market with a serious deficiency was New York City. In the case of only one market, station data would be needed to supplement market data for an analysis. The limitations of this study precludes further inquiry in this area of analysis. It was found that total retail sales for the ADI was the best predictor of potential revenue. An almost linear relationship exists between the two variables. Total television revenue was found to be only slightly less linear than potential revenue with the variable total retail sales. In an attempt to use multiple regression to aid the prediction of potential revenue, other variables included in the model did not supplement the power of total retail sales. Therefore, without the input of per station data, total retail sales alone can be used to predict potential revenue with 96 percent of the variance explained. Conclusions Unlike Yadon's study which suggests that the larger a market gets, the less likely existing stations are to assume the potential revenue available, larger television markets do very well in meeting, and in most cases, ex- ceeding the markets potential. This is perhaps due to the fact that radio is more of a local advertising medium and television relies heavily 79 on national advertising dollars and the benefits of network programming. Another possibility is the scarcity of tele- vision allocations (particularly VHF) making the supply of programming short relative to demand. Radio services also Operate at differing power levels (local, regional, and clear channel) whereas television service is limited to a much more narrow range of power. Perhaps cable tele- vision proliferation in the future will make up for this difference between the two mediums. Although television is classified as a national advertising medium, local market conditions were found to influence, to some extent, the market financial variables. Particularly affected were network compensation and to a lesser extent spot revenue. Known test market cities and some areas of the sun belt, where there is less viewing of television, tend to receive an abnormally high percentage of spot dollars. Atlanta, Miami, Phoenix, and others tend to attract more than their share of spot business based solely on market characteristics considered in this study. Other than these special cases the percent share of spot revenue follows a systematic decline as the markets get smaller. But, even with this slight inconsistency, the potential revenue concept using KPr gives fairly good estimates of how much a particular market should be billing. Discriminant analysis also did a good job in pre- dicting the markets economically eligible for new entry. Although a small number of markets were wrongly predicted 80 by the computer program, these instances, for the most part, could be explained away. If certain qualitative factors could be entered into the program, fewer errors in predic- tion would be expected. This research has made it possible to include several new aspects in decision models as they pertain to constructing new television stations in the tOp fifty mar- kets. The use of total retail sales, potential revenue, and the coefficients produced in the discriminant analysis can add significant new light to television market analyses. Recommendations Based on the results of this study, the author recommends that the top fifty television markets be eval- uated using the potential revenue concept and by applying the formula constructed through the use of discriminant analysis. The constant, Kpr' should be updated annually to insure accuracy, and the amount of viewership should be studied in depth for possible inclusion in general analysis of markets. This exploratory study indicated that audience data influence, considered on a per market basis, produced in- conclusive results relating to the dependent variable. It is therefore recommended that further study be made of the relationship between per station audience data and the financial characteristics of markets where such information is readily available. Such a detailed investigation is 81 recommended and seems necessary to describe Optimum com- petitive standards between stations in markets of varying characteristics, and in different regions of the country. It is further recommended that detailed analysis of hyphenated markets be made to ascertain the reasons why they tend to differ so dramatically from single city markets. Perhaps the task would be made easier if per station finan- cial data were made available. The use of per station ARB data in connection with per station financial data should be studied to determine if the correlation between the two are sufficiently high, so that rating books can be used more extensively in market analysis. Finally, since this study of the top fifty markets only covers a fraction of all the television markets it is suggested that analysis be carried out on the balance of television markets, or at least the next fifty. Perhaps similarities can be spotted among various markets and categorized for simplifying future studies. For instance, the groups may consist of hyphenated markets, growth mar- kets, superconurbation markets, and less urban markets. A SELECTED BIBLIOGRAPHY A SELECTED BIBLIOGRAPHY Books Barnouw, Erik. Tube of Plenty. New York: Oxford University Press, 1975. Besen, Stanley M. The Value of Television Time and the Prospects for New Stations. Santa Monica: Rand Corporation, 1973. Blau, Robert T., Johnson, Rolland C., and Ksobiech, Kenneth J. The Determinants of Television Stations Sales Prices 1968-1973. Bloomington: Indiana University, 1975. Head, Sidney W. Broadcasting in America: A Survey of Television and Radio. 3rd ed. Boston: Houghton Miffin Company, 1976. Kahn, Frank J. Documents of American Broadcasting. New York: Appleton-Century-Crofts, 1968. Kerlinger, Fred N. Foundations of Behavioral Research. 2nd ed. New York: Holt-Rinehart-Winston, 1973. Kerlinger, Fred N. and Pedhazur, Elazar J. Multiple Regression in Behavioral Research. New York: Holt-Rinehart-Winston, 1973. Lichty, Lawrence W. and Topping, Malachi C. American Broadcasting. New York: Hastings House, 1975. Noll, Roger G., Peck, Merton J., and McGowen, John J. Economic Aspects of Television Regulation. Washington, D.C.: Brookings Institution, 1973. Owen, Bruce M., Beebe, Jack H., and Manning, Willard W. Jr. Television Economics. Lexington, Mass.: Heath and Company, 1974. Park, Rolla Edward. Potential Impact of Cable Growth on Television Broadcasting. Santa Monica: Rand Corporation, 1970. 82 83 Park, Rolla Edward, Johnson, Leland L., and Fishman, Barry. Projecting the Growth of Television Broadcasting: Implications for Spectrum Use. Santa Monica: Rand Corp., 1976. Quall, Ward L. and Brown, James A. Broadcast Management. 2nd ed. New York: Hastings House, 1976. Yadon, Robert E. "Financial Behavior of Oklahoma Single Station Markets in 1973." M.A. Thesis, Oklahoma State University, 1975. Articles Besen, Stanley M. and Hanley, Paul J. "Market Size, VHF Allocation, and the Viability of Television Sta- tions," Journal of Industrial Economics, Vol. XXIV (September, 1975), pp. 41-54. Broadcasting (1970). "A Play by Play Retrospective," November 2, 1970, p. 114. Broadcasting (1977). "Lee Sees TI Tuner As Spectrum Saver," December 12, 1977, p. 60. Broadcasting (1978). "Musical Chairs in Alabama," January 9, 1978, p. 37. Broadcasting (1978). "Special Report," January 2, 1978, p. 28-29. Broadcast Management/Engineering (1976). "Satellites: Growth Competitor to Land Lines and Air Freight," October 1976, p. 56. Broadcast Investor (1976). Paul Kagan Associates, Inc. September 1976. Lachenbruch, David. "The Three Billion Dollar Gamble," TV Guide, November 1, 1975, p. 5. Lee, Robert E. "The Drop In PrOposal: Let's Drop the Matter and Give UHF a Chance," Television/Radio Age, March 15, 1976, p. 57. Webbink, Douglas W. "Regulation, Profits and Entry in the Television Broadcasting Industry," Journal of In- dustrial Economics, Vol. XXII (September 1973), pp. 167-176. APPENDICES APPENDIX A THE TOP FIFTY TELEVISION MARKETS IN 1975 APPENDIX A THE TOP FIFTY TELEVISION MARKETS IN 1975 oxoooxlmmnwmla H 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. New York Los Angeles Chicago Philadelphia Boston San Francisco Detroit Washington, D.C. Cleveland Pittsburgh Dallas-Ft. Worth St. Louis Minneapolis-St. Paul Houston Miami Atlanta Tampa-St. Petersburgh Seattle-Tacoma Baltimore Indianapolis Hartford-New Haven-New Britain- Waterbury Milwaukee Kansas City Portland, Oregon Sacramento-Stockton Cincinnati Buffalo Denver Providence Nashville 84 47. 48. 49. 50. 85 San Diego Columbus Charlotte Memphis New Orleans Greenville—Spartanburg-Ashvil1e Phoenix Louisville Grand Rapids-Kalamazoo-Battle Creek Dayton Oklahoma City Charleston-Huntington-Ashland Albany-Schenectady-Troy Orlando-Daytona San Antonio Harrisburg-Lancaster-York- Lebanon Wilkes Barre-Scranton Norfork-Portsmouth-Newport News- Hampton Syracuse Salt Lake City APPENDIX B INDIVIDUAL MARKET INDEXES O O O O O O O O OkOCDQmU'Ioh-UJNH H 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Market NYC LA CHI PHIL BOS SF DET WASH CLE PIT DAL ST L MINN HOU MIA ATL TAM SEA BAL IND HART MIL KC PORT SAC CIN BUFF DEN PROV NASH APPENDIX B INDIVIDUAL MARKET .51 .49 .50 .57 .51 .53 .42 .44 .59 .52 .54 .46 .48 .45 .42 .51 .61 .55 .37 .37 1971 .41 .56 .49 .43 .48 .54 .47 .42 .46 .47 .51 .44 .49 .58 .52 .55 .41 .44 .59 .47 .52 .45 .46 .46 .45 .52 .70 .55 .34 .39 86 .57 .56 .43 .42 .57 .49 .48 .47 .49 .45 .50 .51 .69 .63 .37 .38 INDEXES 1973 .43 .57 .51 .42 .52 .58 .48 .43 .45 .47 .47 .45 .48 .62 .57 .60 .40 .40 .56 .49 .47 .48 .46 .40 .45 .48 .77 .57 .35 .38 .42 .50 .44 .45 .49 .59 .54 .55 .37 .40 .51 .48 .44 .45 .50 .39 .44 .45 .72 .61 .33 .38 .40 .47 .48 .71 .61 .33 .35 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. Market 8 DI COL CHAR MEM N O GV/S PHO LVIL GR/K DAY OK CHAS ALB ORL S A HAR WB/S NOR SYR SLC .45 .40 .65 .35 .57 .49 .37 .47 .43 .37 .44 .37 .41 .28 .30 .45 .52 .41 .48 .40 .67 .36 .58 .49 .38 .49 .43 .36 .44 .38 .44 .29 .32 .44 .44 .45 .53 .44 .39 .65 .30 .61 .47 .41 .48 .45 .37 .47 .37 .48 .31 .29 .40 .52 .42 .48 .33 .44 .35 .44 .28 .28 .38 .43 .45 .38 .37 .57 .47 .37 .45 .45 .36 .48 .37 .45 .29 .31 .39 .41 .45 APPENDIX C POTENTIAL REVENUES AND EFFICIENCY INDEXES FOR THE TOP 50 MARKETS 88 Nmm. vm>.mv NHo.H vmh.mm omm. How.mm mHN.H wah.mm vmo.a www.mm hmm. omm.mm Hho.a Nam.moa mNN.H mmH.mNH mam. mmw.OHN mbma wmm. mam.m¢ hmm. wa.mv Ham. Hmv.mm NHN.H mov.mm Hmo.a avo.nm mum. hov.hh hmo.H HHO.NCH mmH.H vmm.maa mmw. th.oom whoa mBmMmflz om mmm. mom.Hv mom. mmH.mv wmo.H Mbh.om mmm.a owm.om moa.a owv.mm mam. www.mn mho.H mom.mm wom.H omH.HHH mom. mob.omH mhmH omm. mmm.mm amm. mvm.mv mmm. mmv.¢v Hom.H ham.hv mmo.H mam.Hm mam. NhN.on mmo.H bmo.mm hom.H mam.moa mmm. mov.¢ma mea Ham. vam.vm mom. mHN.mm moo.H oom.mv omH.H www.mv Hmo.H www.mv mmm. oma.mm Nmo.H www.mh mmH.H www.mm mmm. mmh.mma Head moo.H wmm.mm Ohm. mmn.vm mmm. Hmw.vv ovH.H mmh.mv mmo.H mam.mv Hem. NmH.Nw ovH.H HmH.mh mmH.H mmm.mm owm. mnm.mma onma AmCOHHHaEm on woco>om Hoflucouomv U xHQmemfl mOB mmB m0 NMQZH MUZHHUHmmm QZ< szm>mm AmHU .U.Q .coumcflcmmz pflouumo OUmHUCMHh Cmm coumom naroaoonaard ommoncu mmaomc< moq xuow 3oz porno: mo we no we mo vo mo No Ho 89 0mm. wam.mm mvH.H www.mm mum. 5mm.m~ mmm. ovm.hm hma.a mmm.Hm hmH.H Hmm.mm mHN.H vab.mm Hoo.H mma.vm Hoo.H moo.mm mqo.H me.Hv me.H www.mm mhma mmo.a Nam.vm moa.H mmh.mm mew. mmm.sm Ham. mmo.hm boa .H oav.mm mmH.H mwm.Hm th.H th.mm owo.H mmv.om Ham. omH.Hm Hum. 5mm.mm mmo.H owm.om «baa mNo.H mwm.m~ va.H vwm.am va. Nam.wm hmm. Hmo.mm H5N.H bmm.vm hmH.H moo.hm mHM.H hom.vm mmo.H mmw.hm mmm. mmm.mm voo.H mHH.mm moo.H moa.om mead mHo.H mmv.HN th.H hmm.om owm. Hma.vm Hem. www.ma mmH.H www.mm omH.H wmo.vm va.H NHw.N~ mmo.a Hhm.mm mmm. Hom.hm Hoo.H vma.mm mam. hom.mm tha mam. mov.wH mom.H Ham.ma mom. mom.mH 0mm. Mbm.mH HmH.H omm.wH wHH.H mom.om hm~.H wwm.ma wHH.H mom.HN mmm. www.mm Hmo.H Hmo.hm voo.H www.mm Head NOH.H mmm.ma mmN.H mon.ma mmm. mnm.om vom. me.vH mNH.H mnm.ma mmo.H vvm.ma NHN.H mmm.wa hvo.H www.mm mvo.a mma.mm OHH.H mmm.hm mam. Nmm.mm onma maaoomcmwch muosnuamm mEoomalmHuummm muonmumuom .um IomEoe mucmHu¢ HEMHZ coumoom Home .um ImfiaommmccHz masoa .um nunos .umummaaoo ransonupae porno: om ma ma NH ma ma vH ma NH HH 0H 90 HNN.H mnm.ma vmh. www.ma van. www.ma HHM.H vmv.vm mmm.a Hmo.mH Nvo.H mmm.HN NHo.H mam.HN 9mm. mvv.mm meo.a mom.mm Hmm. mmm.mm mmm. Nam.mm mhmfi mvN.H mh0.mH mmw. mmm.ma mas. mnm.mH mmm.a mvm.HN mmm.a mmo.mH mum. vom.om Aha. hma.om vmm. vvm.om Hmo.H mmm.HN Nmm. moo.NN mmm. ovm.¢m vbma th.H www.ma mom. mmm.va vmw. mvm.wa vom.H Hmo.om omw.H mmw.ma mmo.H mmm.ma mmm. Hmo.mH omm. omm.mH omm. Hmo.om hoo.H hmm.om mam. mwm.HN mnma omo.H Hmw.NH has. mmn.ma mun. mom.ma mom.a Nom.ma mmv.H owm.hH mmo.H NHV.>H mmo.H mhm.bH mmm. mmm.wH NNO.H www.ma «mm. www.ma ham. mom.om mhma Hmm. mhm.oa Nvm. mvv.HH mmn. hmv.va mmH.H OHH.vH 5mm.H oav.mH NHH.H HmH.mH 5mm. mmm.ma 0mm. onw.ma Ham. mmm.ma vmm. Nmm.wH mOH.H hoo.hH Hbma mmm. Nvm.HH was. wmm.HH mom. th.vH me.H mnm.ma NHM.H NmH.mH mmo.H vmm.mH Ham. Hmm.mH omm. omN.vH mmo.H vmm.ma who. hum.ha me.H www.mH osmH omoao com OHHfl>£mmz oococa>onm Hm>cmo Oammwsm auMCCHocHU oucoEmuomm pcmenom xuno momcmm wmxom3awz co>mm 302 readmuamm porno: Hm om mm mm mm mm mm vm mm mm Hm 91 aha. www.ma mwm. moa.ma vow. avm.mH HHo.H omc.mH HMN.H mmv.ha mow. moa.va c¢N.H vom.ma mam. Hmo.ma omo.a mom.ma mo~.H mHH.hH mhma ovo.H maa.ma mhm. omw.wa cam. www.ma mmm. ovm.va mmN.H oom.ma NNm. wmo.ma hmm.a moh.ma va. mnm.ma Nam. www.ma mmo.H Hmm.oH «mmm vmm. www.ma oao.a www.ma Hmm. Hmm.vH mmm. omm.ma th.H Hmm.va mmm. Hmm.va Nmm.H omm.NH mmm. Nmm.mH mmm. NOH.vH NNH.H Hmm.vH mbma omm. who.NH omm. mmN.NH NNm. mmm.ma mmo.a MNN.NH hMN.H moo.ma Now. hmm.oa mmm.H Hmo.HH mmw. mom.va avo.H wam.NH mvH.H omv.ma whoa omm. HNN.OH vvo.H hbv.oa Hmm. mmm.aa Nmo.H oom.oa th.H va.HH mhn. va.m va.a Hmm.m mwm. Hmm.HH ovo.a www.0H voo.a Hmo.ma HhmH hmm. mon.m moo.H mmm.oa com. mnm.ma mqo.H vmo.m vNN.H HNO.HH th. mmo.m mmm.a NHN.m mum. Hmo.NH Hum. m¢N.0H hmo.H hao.ma onma noao pacemaxo couwmo xoouu mauumm OONmEmHMM Impfimmm Ocmuo maan>masoq choocm maaa>on< Imusncmunmmm IwHHn>cmouw mommauo smz mnsmfimz ouuoanono monEOHou porno: Hv ov mm mm hm mm mm vm mm mm 92 ohm. www.ma 5mm. www.ma omm. mhm.NH hum. HBQ.HH omo. www.ma hum. th.va mob. th.mH hqo.a Hmo.ma awn. mow.NH mhma mom. moo.ma mmm. mam.aa 0mm. www.ma ham. va.HH mam. mmo.mH Ohm. va.ma mmh. www.ma mom. mHN.mH man. mmm.HH whoa haw. Hmm.NH NOH.H hum.m mmm. mow.HH omw. mam.oa mmm. www.ma Hmo.H NHH.HH th. Nov.ma voo.H mmw.HH ebb. Hmm.OH mnma mum. hoa.oa mom. omm.¢ Now. bmv.oa vmm. mmm.m mmm. mmm.mH hvo.a oma.oa ohm. omv.HH hwa. omv.HH can. th.m Mnoa vwm. mvo.¢ mvm. hov.m mvm. mow.m mam. mmo.h Maw. me.NH mom. who.m wmm. hmo.m vvm. mmm.oa mom. ovv.m Huma «mm. mmo.m mmm. HNH.OH 5mm. Nov.m omm. mmv.h mom. me.NH vhm. Hom.m mmh. mmo.m mmm. mwh.oa Nah. hma.m Chad nuao mama papm wmoomuwm msoz uHOQSOz (cuooamuuom IxHOmnoz coucmuom nonumm mmxawz cocmnmq (kummocm41xuow (muonmnuumm Oficoucm com occmHuo ncmnaa coumcwucom Icoumoaumco porno: om mv mv 5v ow mv vv me me APPENDIX D ANALYSIS OF VARIANCE FOR THE UPPER AND LOWER 25 MARKETS IN THE TOP FIFTY MARKETS PER YEAR APPENDIX D ANALYSIS OF VARIANCE FOR THE UPPER AND LOWER 25 MARKETS IN THE TOP FIFTY MARKETS PER YEAR Sum of Mean . Year Source df Squares Squares F ratio 1970 Between Groups l 3.13:? 3.13:3 6.70 p < .05 Within Groups 48 2.24_7 4.67 Total 49 2.56 1971 Between Groups 1 6.35:: 6.35:3 2.21 (n.s.) Within Groups 48 2.43_7 6.38 Total 49 3.06 1972 Between Groups 1 8.18:? 8.18:3 1.44 (n.s.) Within Groups 48 3.37__7 7.02 Total 49 3.47 1973 Between Groups 1 8.15:? 8.15:3 1.52 (n.s.) Within Groups 48 3.10_7 6.46 Total 49 3.92 1974 Between Groups 1 8.47:? 8.47:3 1.58 (n.s.) Within Groups 48 2.75_7 5.73 Total 49 3.60 1975 Between Groups l 8.71:; 8.71:; 1.32 (n.s.) Within Groups 48 3.16 6. 8 Total 49 3.24 93 APPENDIX E LIST OF MARKETS QUALIFYING FOR NEW ENTRY BY YEAR APPENDIX E LIST OF MARKETS QUALIFYING FOR NEW ENTRY BY YEAR 1970 N=12 New York City Los Angeles *Chicago San Francisco Detroit Washington, D.C. Baltimore Indianapolis *Kansas City *Buffalo Phoenix Dayton 1971 N=17 New York City Los Angeles San Francisco Detroit Cleveland *Atlanta *Tampa—St. Petersburg Seattle-Tacoma Hartford-New Haven Kansas City San Diego Charlotte Phoenix *1ouisville Dayton San Antonio Norfolk 1972 N=3 New York City *Los Angeles Charlotte 1973 N=8 Los Angeles San Francisco Detroit Pittsburgh Dallas-Ft. Worth Sacramento Charlotte Phoenix 1974 N=12 *New York City Los Angeles Detroit Dallas-Ft. Worth Houston Indianapolis *Sacramento Buffalo Charlotte Phoenix Harrisburgh-York Norfork-Portsmouth 1975 N=9 New York City Los Angeles *San Francisco *Detroit Houston *Miami Indianapolis Charlotte Norfork-Portsmouth * Signifies actual entry 94 || 23 TE UNIV. LIBRQ S RIE ||| ||||||||||||||||| 1 1 ES $21 030 MICHIGAN s |||||||||||||||||| 3129