This research is a longitudinal and comparative study on the merger/acquisition(M&A) strategies by cable operators and telephone companies in the multimedia context. The unit of analysis is a cable or telephone company for the first part (i.e. H
Research Population and Census Data
The population of this study is all mergers and acquisitions completed by U.S.-based cable systems and telephone companies from 1984 through 1999. All identified mergers/acquisitions were coded as multimedia-related or non-multimedia-related. Because there are multiple definitions for multimedia industry, the selection of multimedia-related industries was based on several government documents and previous studies that focused on the issue of multimedia, information, or media convergence.
The major documents that provided detailed listing of multimedia-related sectors included the following:
(1) a study on media conglomerates (Albarran & Dimmick,1996);
(2) an examination ofthe convergence of information and communications(Chakravarthy & Rodan,1999);
(3) two government reports released by the Department ofCommerce on information technology industries (DOC,1999,2000);
(4) the definition of information industry based the North American Industry Classification System (NArCS) released by the U.S. Census Bureau.7
Other similar studies (Collis, Band & Bradley, 1997; Colombo & Garrone, 1998; Jamison, 1999) were also taken into consideration in defining the multimedia market boundary, though these studies did not provide detailed industry lists.
Because each document took a slightly different approach to the definition of multimedia industry, altogether more than 56 individual sectors were documented that were one way or another related to information or multimedia (Table 3-1). In order to make the data more consistent, we decided that each sector must be mentioned at least by two of the four principal data sources in order to be treated as multimedia-related.Several sectors that are more related to information technology, such as semiconductors and circuit boards, and some sectors, such as greeting cards and ad agencies, were decided as not consistent with this study and were not included in the sample. This resulted in a total of 42 sectors as the multimedia-related sectors for this study. All the mergers and acquisitions completed by cable and telephone companies during the 16 years under investigation were coded based on the selected multimedia-related sectors, using SIC codes.
Table 3-1. List of selected multimedia-related industries.
…………………………………..
Mergers and acquisitions coded as multimedia-related were further classified into seven categories, mostly based on the SIC codes. Table 3-2 summarizes the grouping results. Although the categorization is arbitrary to some extent, the largely similar SIC codes in each category suggest that the grouping is adequate.
Data sources and coding scheme
The major archival data source is the merger/acquisition database (SDC Platinum) complied by Securities Data Co. The database includes all corporate transactions. Private as well as public ones. that involve at least 5% of the ownership of a company with transaction valued at $1 million or more. Transactions involving at least 5% of ownership but without available data of transaction value are also included. Starting from 1992. all transactions ofany value are included in the database. provided that the transactions involve at least 5% of ownership.
Merger/acquisition data recorded for analysis included: (1) year in which the
transaction was completed. (2) the acquiring fmu's name. business description. and SIC code, (3) the target firm' s name. nationality. business description. and SIC code. and (4) transaction value if available. Other data such as corporate revenues. market size, and market growth rates, were collected from other existing databases such as government documents and corporate data compiled by private for-profit organizations. The database industry Norms and Key Business Ratios compiled by Dun & Bradstreet Corporation is the major source for industry-level financial data. Two databases of Ward's Business Directory published by Gale Group and Directory of Corporate Affiliations by National Register Publishing are the major sources ofcorporate-level financial data.
Variable Operationalization and Measurement
The goal of this study is to explore the merger/acquisition strategies by the cable systems and telephone companies, and to examine the impact of numerous factors on merger/acquisition decisions. This study proposed several factors that might have impact on a company's merger/acquisition decision. These included changes in regulatory environment, business status ofthe acquiring firms, attractiveness oftarget markets, the relationships between the acquiring and target industries, and the impact of prior merger patterns. Table 3-3 presents the summary of operational definitions and measurement.
Table 3-2. Classification of multimedia sectors.
………………………………………………….
Regulatory environment
This study used dummy variables to capture the impact of regulatory environment. Since the divestiture ofAT&T and passage ofthe 1984 Cable Act in 1984, telecommunications concerning the telephone and cable industries have underwent several amendments at different points in time. Table 2
In addition, because the cable and telephone industries are subject to different regulatory framework, another dummy variable (ACQUIRE) was used to capture the different regulatory schemes. This dummy also served to represent the different market nature and technological characteristics between the cable and telephone industries.
Business status of acquiring firms
This study used two interval variables to reflect two aspects ofbusiness status of the acquiring firms. First, three-year average ofrevenue gro\Vth rate prior to merger completion (SALEG) represented the history ofgrowth prospect of the acquiring firm in terms ofsales volume. Second, three-year average of net income growth rate prior to merger completion (INCOG) represented the ability ofthe acquiring firm to generate profit.
Table 3-3. Operational definitions and measurement ofvariables.
………………………………….
It should be noted that these two aspects are not necessarily highly related~ and both sales and income stream could play an important role in managerial decision for corporate expansion (Ansoff,1988). Sales revenue does not factor in sales costs or other administrative expenses. Thus, corporate expansion via a merger or acquisition may help the acquirer to increase sales revenues, but it does not necessarily also increase profit for the company. A manager faced with slow sales growth may have higher incentive to pursue business combinations because it helps to pump up sales volume quickly.
On the other hand~ net income factors in operating, administrative, and interest
expenses. The prospect of net income growth (i.e., the bottom line) of a company is often what stockholders pay close attention to, as it commonly determines the expected dividends (Redmond & Trager, 1998, chapter 9). Thus. when a company is confronted with slow income growth or even losses, pressures from stockholders may also be an important reason for a manager's decision to pursue mergers.
Attractiveness of merger target industries
Three interval variables-revenue growth rate, profit growth rate, and return on
assets-were used to capture the different aspects of attractiveness ofa merger target
industry. The variable SALEGRO, which represented three-year average of annual
growth rate prior to merger completion~was used to capture the growth prospect of a target market. The variable "t\.TETGROA~ which represented three-year average of annual income growth rate prior to merger completion. was used to capture potential of income generation of a target industry. The third variable ROAAVG represented three-year average of return on assets (ROA) prior to merger completion, and served to capture the profitability of a target industry. ROA is commonly used as a key measure of management productivity, an indicator ofhow efficiently a company’s assets are being utilized (SobeL 1993). To the extent that merger and acquisition is transaction of corporate assets, the potential return on assets would be a key indicator of market attractiveness for the acquiring firm.
Acquirer-target relatedness
This study focused on three aspects of relationship between the industry of the acquiring firm and the target industry: customer relatedness, vertical relatedness, and technological relatedness.
Customer relatedness. First, customer relatedness (CUSTOME) was defined as the degree to which the products/services provided by the acquiring industry are functionally similar or complementary to those provided by a certain target sector. For example, cable television and broadcast television both have the same customer source (i.e., audiences as well as advertisers). However. it should be noted that advertising revenue is the major income source for broadcasters, whereas subscription fee is the major income source for cable systems. Video production studios were also considered to have the same customer base as cable systems. Likewise, plain old telephone and cellular phones are functionally similar and complementary to each other. This study used a dummy variable to measure customer relatedness.
The measurement ofcustomer relatedness was based on common observation
The product provided by conventional cable television system is video delivery services, whereas the product oftelephone companies is voice transmission services. Based on this rule of thumb, target sectors such as broadcast television and other video markets were considered related to the cable television industry. Alternatively, voice transmission services such as wireless telephony (e.g. paging and cellular) had a high degree of similarity from consumers· viewpoint. Other sectors were treated as unrelated.
Vertical relatedness. Second, the vertical relatedness was defined as the conventional sell-buyer relationship. This variable was divided into two sub-units: forward integration (FORWARD) and backward (BACKWAR) integration. The construction of relatedness index was based the input-output accounts compiled by the
Department of Commerce (DOC, 1998).
The input-output accounts provide very useful data in dollar term of (1) input that a certain industry requires from other industries for its production to final users and (2) output that an industry makes for other industries’ output. The index reflects the degree of“resource dependency” between the acquiring and target firms (Klavans, 1990). For the present study, the former index represented the degree of backward vertical relatedness; the latter represented the forward vertical relatedness.
Note that the utilized data (i.e., "industry-by-commodity total requirements
coefficients")provide only the indices for the 15 largest producing industries for each industry's production. The remaining is categorized as "'all others." However. for a majority of industries, the top 15 industries commonly account for over 80% of the total input required by an industry for its production. The missing data should not cause considerable biases. The value zero was given for those missing data.
Technological relatedness. The third index was defined as the degree of
relatedness between two industries in terms ofthe primary technology for goods or
services production. The technological relatedness (SICSIM) index was adopted from the SIC-based "concentric index" developed by Caves. Porter, Spence, and Scott (1980, Chapter 8). This concept was also used by other studies on corporate diversification(Bergh. 1998; Palepu, 1985; Robin & Wiersem,1995). Based on this concept., the more similar the SIC codes ofthe acquiring and target industries, the more related they are.
Impact of prior mergers on subsequent ones
According to the isomorphism theory discussed earlier. a firm tends to follow
what the other companies with similar characteristics are doing. To test the theory, this study uses two variables to represent the behavior of other acquiring frrms within the same core industry-that is~ the total number (CUMNUMB) and amount oftransaction value (CUMNUMB) of mergers and acquisitions in each target sector that have been completed in the past three years by companies within the same industry (i.e., by cable systems or by telephone companies).
It is worth noting that we expected number of transactions to be a better indicator of merger/acquisition momentum than value of transactions for two reasons. First. The magnitude of transaction value is likely to be affected by some large-scale transactions.
In this situation, the trading activity might be overestimated. On the other hand, the use of transaction number can factor out the influence ofsome large-scale deals. It can better capture the general merger momentum because some smaller-size companies could also be active in business combination though the trading level in dollar term might be small.
Research Design and Variable Validity
To ensure the validity of the results generated from this study, variable validity must be checked. Because the variables such as revenue and market growth have face validity and reliability, the issue of construct invalidity should not arise.
However, a major concern may arise with the validity of the indexing schemes developed to represent the degree ofacquirer-target relatedness in this study. As mentioned earlier, the degree of customer relatedness was based on the observation of functional similarity and complementarity of products/services. Although there is no standard criterion for the measurement ofrelatedness, the definition was considered as consistent with previous studies on the adoption ofnew media/communications technologies (Atkin, Jeffres, & Neuendorf: 1998; Dupagne, 1999; Jeffres & Atkin. 1996; Perse & Dunn, 1998). That is, consumers tend to adopt those new media/communications technologies that provide utilities similar to those ofthe existing ones they are using. Because conswner utilities and needs are important concepts in marketing and management disciplines. product similarity should serve as an important indicator of relatedness for managers. Thus, the construction of index of customer relatedness based on observation of product similarity was believed to have face validity.
In addition. operational defmitions of several variables in this study were based on the U.S. SIC system-for example. the classification of industrial sectors, the classification as diversification or horizontal expansion, and the degree of technological relatedness. The SIC system was developed to facilitate collection of data for economic analysis. It employs a set ofreporting standards that have evolved over time based on a variety ofconsiderations ranging from similarities in materials to product-market linkages. Some scholars have challenged the validity of using SIC to reflect the interrelationship among industries and attempted to develop alternative methods for that matter (Rumelt. 1982; Robins & Wiersema, 1995; Teece, 1982). Although the use of the SIC system provides only limited information on relationships among industries, it has some attractive features that makes its use popular in industry analysis (Hoskisson, Hitt,
Johnson. & Moesel,1993). Because the data are classified according to standard
categories! they have the merit ofconsistency! which helps to make research replicable and cumulative. This is especially important for longitudinal studies. Moreover. It allows researchers to use the broad range of data that are reported using the SIC system. Under these circumstances! the SIC system still serves as a useful basis for this study.
Statistical Procedures
Two statistical methods were used to test the hypotheses. Logistic regression(logit) was performed to test hypotheses
For hypotheses two, three, and four, analysis ofcovariance (ANCOVA) was
performed. Similar to regression, ANCOVA procedure provides straightforward
statistical tests and a wide range ofdiagnostics. But this procedure is appropriate when the dependent variables are quantitative ones, and when the independent variables consist of qualitative and quantitative ones (Agresti & Finlay! 1997! chapter 13). The flexibility of ANCOVA allows us to accommodate the categorical variables such as time period and industrial categories (i.e." cable and telephone).
Hypotheses la through 1d
Hypotheses la through Id stated that cable systems and telephone companies were more Likely to pursue diversifying mergers/acquisitions and international expansion in the 1990s than the 1980s. In addition, companies faced with slower revenue or net income growth were more likely to make diversifYing acquisitions and internationalization.
The binary dependent variable was defined as whether the merger transaction was a diversifying (= 1) or non-diversifying (= 0) one, or whether the merger transaction was a domestic (= 0) or international (= 1) one. The independent variables included ACQUIRE (core business of acquiring firms), PERIODI and PERIOD2 (three different time periods), INCOG (three-year net income growth rate ofacquiring firms), and SALEG (three-year revenue growth rate of acquiring firms).
The model consisting of only dummy variables ACQUIRE, PERIOD 1. and PERlOD2 was tested first. and served as a baseline modeL. Then the model with the addition ofquantitative independent variables SALEG and INCOG was tested for comparison. This procedure allowed us to examine the improvement of explanatory power of the independent variables INCOG and SALEG.
Hypotheses
Hypotheses
diversification were more Likely to acquire target firms in the industries with higher growth rate and profitability. To test these hypotheses, ANCOVA was performed.
The methodology followed that used by KIavans (1990) and Pfeffer and Salancik (1978). The dependent variable was the ratio of the number oftransactions (and total value of transactions) ofmergers/acquisitions cable or telephone companies made into individual industries within a single year to the total mergers completed by cable or telephone companies within the same year. To a large extent, the percentage based on transaction value represented the degree ofresource allocation (i.e., capital investment) by the cable and telephone industries for diversifying mergers in each ofthe non-core industrial sectors. The percentage based on transaction number represented the occurrence of merger/acquisition in a certain target industry. Total number (transaction number ratio) and total value (transaction value ratio) were tested separately. The independent variables were ACQUIRE. PERIOOl. PERI002.SALEGRO (revenue growth rate). NETGROA (income growth rate), and ROAAVG (profitability).
Similar to the previous logit procedure~ the model consisting ofonly dummies ACQUIRE, PERIODl and PERIOD2 was tested first and served as a baseline model. This procedure allowed us to examine the degree ofvariation accounted for by SALEGRO, NETGROA, and ROAAvG.
Hypotheses
Hy-potheses
Hypothesis four
Hypothesis four stated that cable and telephone companies pursuing
diversification were more likely to follow the same patterns ofacquisitions that had been previously repeated by other acquiring flffilS within the same industry. To test this hypothesis, the same statistical procedure was performed, and the same dependent variables were tested separately. Independent variables were ACQUIRE, PERIOD1, PERIOD2, CUMVALU (three-year cumulative transaction value) and CUMNUMB(three-year cumulative transaction number). The baseline model versus expanded model procedure was performed for comparison purposes.