Stakeholder Participation and FCC Broadcast Policy Decision Making

Research Report Prepared with a Grant from the National Association of Broadcasters

Philip M. Napoli

Assistant Professor

Graduate School of Business

Fordham University

113 W. 60th St.

New York, NY 10023

212-636-6196

pnapoli@fordham.edu

Abstract

This paper investigates whether the volume, source, and position of formal comments filed by various stakeholder groups have had an effect on the FCC's broadcast policy decision outcomes. A sample of broadcast policy decisions and their associated stakeholder comments were analyzed for the 1992 through 1997 time period. Comments were gathered and analyzed via the Record Image Processing System available at the Federal Communications Commission. The results indicate that the overall distribution of stakeholder comments (in terms of favoring versus opposing policy proposals) is not related to the FCC's decision outcomes. However, when the stakeholder comments are separated into categories (broadcast industry, individuals, competing industry segments, other), the results indicate that both broadcast industry and competing industry comments are significantly related to the FCC's decision outcomes. Overall, the multivariate logistic regression model explains over 40 percent of the variance in the dependent variable.

Introduction

There is a long history of research devoted to the decision making of the Federal Communications Commission and the potential influence of various external stakeholders on the communications policymaking process (e.g., Braun, 1994; Canon, 1969; Cohen, 1986; Edelman, 1950; Gormley, 1979; Krasnow, Longley, & Terry, 1982; Lichty, 1962; Napoli, 1997, 1998b). ⁽¹⁾ Theories of regulatory behavior have frequently focused on the influence potential of a single stakeholder group. For example, theories of "congressional dominance" (Weingast & Moran, 1983) or "industry capture" (e.g., Berner, 1976; Cole & Oettinger, 1978; LeDuc, 1973; Linker, 1983; Schwartz, 1959) have often been applied to the FCC; however, such theoretical perspectives have frequently come under criticism for oversimplifying the dynamics of the policymaking process and for neglecting the multitude of stakeholders with diverse and potentially conflicting interests that participate in this process (see Teske, 1990; Napoli, forthcoming; Williams, R.J., 1976; Woolley, 1993; Symons, 1989).

A more sophisticated approach to understanding the dynamics of stakeholder influence on the communications policymaking process may come from insights developed primarily within economics and organizational behavior regarding the relationship between "principals" and "agents" in any organizational context (see Alchian & Demsetz, 1972). According to principal-agent theory (or agency theory, as it is commonly known), an organization such as the FCC can be conceptualized as an agent serving a number of principals (or overseers), including the general public, Congress, the White House, and the regulated industries (see Levine & Forrence, 1990; Weingast, 1989; see also Krasnow, et al., 1982). That is, the FCC is meant to provide services to each of these categories of stakeholders. According to the principal-agent approach, the degree to which the FCC provides such services may be dependent upon the degree to which each of these principals is "monitoring" the Commission's activities. Lax monitoring on the part of the principals (often referred to as "slack") provides greater opportunity for self-interested behavior (often labeled "shirking") on the part of the agent. Intensive monitoring, on the other hand, increases the likelihood of the agent complying with the principal's wishes (Alchian & Demsetz, 1972; Jensen & Meckling, 1976). Thus, as Crotts and Mead (1979) state simply, the FCC is most likely to respond to whomever is paying attention. Agency theory also emphasizes that the intensity of stakeholder monitoring -- and hence, influence -- is not a constant. Instead, it is a function of "monitoring costs"--the costs associated with effectively monitoring an agent in order to minimize self-interested behavior (see Alchian & Demsetz, 1972; Jensen & Meckling, 1976). This theoretical approach has provided useful insights into the behavior of other regulatory agencies (see Weingast, 1989; Weingast & Moran, 1983), but has yet to be applied to the analysis of the FCC.

In an effort to apply the agency theory perspective to the behavior of the FCC, this study focuses on the role of formal comments filed by various stakeholder groups in regards to the Commission's broadcast policymaking activity. Within the agency theory context, these comments serve as a measure of both the regulatory preferences of the various stakeholder groups, and the intensity with which these groups are monitoring the behavior of the Commission. Thus, an increased volume of comments in regards to a particular policy issue can be seen as an indicator of increased monitoring on the part of the stakeholder group. The results will provide an indication of whether the FCC is indeed responsive to the monitoring of its principals, and, if so, to which principals it is most responsive.

Research Questions

The following research questions guided this investigation into the relationship between stakeholder participation and policy outcomes. The research questions focus on the general categories of stakeholders most frequently identified in the literature on the communications policymaking process. However, given that this research focused on formal comments as a measure of monitoring, research questions regarding the potential influence of political stakeholders such as Congress and the White House could not be asked or investigated, given that their monitoring and/or influence efforts are not manifested in formal comments.

Research Question #1: Is there a relationship between the overall distribution of publicly filed comments and the FCC's broadcast policy decision making?

Research Question #2: Is there a relationship between the quantity and position of comments filed by the general public and the FCC's broadcast policy decision making?

Research Question #3: Is there a relationship between the quantity and position of comments filed by public interest groups and the FCC's broadcast policy decision making?

Research Question #4: Is there a relationship between the quantity and position of comments filed by the broadcast industry and the FCC's broadcast policy decision making?

Research Question #5: Is there a relationship between the quantity and position of comments filed by competing industry segments and the FCC's broadcast policy decision making?

Methodology

In order to empirically test whether the intensity of stakeholder monitoring is related to FCC broadcast policy decision outcomes, a sample of such decisions was generated. For each decision in the sample, all of the formal comments available on the FCC's Record Image Processing System (RIPS) were content analyzed in terms of the source (i.e., stakeholder category) and the position of the comments (i.e., for or against the policy proposal). This section first discusses how the sample of policy decisions was obtained and analyzed. Next, it describes how the comments were obtained and analyzed. Finally, it describes the statistical procedures used to analyze the data.

Policy Decisions

Given that stakeholder comments are only available electronically for dockets dating back to 1992, January of 1992 was the start date at which FCC broadcast policy decisions were sampled. Using the LEXIS "FCC" database, the following keyword search was used for each calendar year from 1992 through 1997: "ACTION(ORDER) and OPINION(BROADCAST)."This search string produced every FCC decision classified as either an "order," "report and order," or "memorandum opinion and order" containing the word "broadcast" in the text of the decision. ⁽²⁾

These decisions were then examined to determine whether they met the following criteria: (a) Did the decision pertain to terrestrial broadcasting? (b) Did the decision modify, eliminate, or enact a rule or policy pertaining to the terrestrial broadcast industry or the role or function of the FCC in relation to the broadcast industry? A decision needed to meet both criteria to be included in the final data set. Many of the cases captured by the search string did not meet these criteria. In many cases, the term "broadcast" appeared in the text of a decision involving another industry (e.g., satellite, cable, cellular). Cases of this type were excluded. Also, many of the cases produced by the search string involved modifications to the table of allotments, broadcast license applications, license renewals, and forfeitures. These were excluded in order to maintain the focus on policy decisions. A total of 34 decisions meeting the above criteria were obtained for the 1992-1997 time period. From these, 19 were randomly selected for content analysis. ⁽³⁾

The analysis of the policy decisions involved three stages. The first stage involved identifying each of the individual policy issues being addressed within the decision. That is, within any decision document, the Commission was deciding on anywhere from one to over 20 individual policy issues. Within a Notice of Proposed Rulemaking (NPRM), the Commission typically presents policy proposals pertaining to a range of related issues. For instance, a 1997 decision regarding amending the rules pertaining to the process of obtaining permits for making changes in broadcast facilities (Federal Communications Commission, 1997) contained 13 separate policy issues, including: whether to increase the permitted variance in the location of the antenna radiation center for FM and television stations; whether to permit broadcast stations to change from commercial to non-commercial status on a license application rather than a construction permit application; and whether to allow commercial FM stations to increase their effective radiated power (ERP) without requiring a construction permit.

Commenters typically focus their comments on one or more (sometimes all) of the individual issues presented in the NPRM. The resulting order then typically summarizes each policy issue, the Commission's specific policy proposal (sometimes more than one proposal is offered -- these were coded as separate issues), selected comments from interested stakeholders, and the Commission's ultimate decision regarding each individual issue. Thus, it is important to recognize that it is the decisions made at the level of these individual issues that comprise the main unit of analysis for this study. Thus, within this sample of 19 decisions, 89 individual policy issues were identified.

Next, each individual policy issue was coded in terms of whether the Commission's policy proposal was adopted or not adopted. This determination was easily made through a reading of the order. Finally, each individual decision issue was coded according to the type of issue presented, to investigate the possibility that monitoring intensity from the various potential sources may vary with the type of issue involved. For example, highly technical issues may not attract the attention of the general public or public interest groups, whereas their interest may be more acute for issues with larger political or social ramifications, such as content or ownership regulations (McGregor, 1986; McQuail, 1992; Stigler, 1971, pp. 10-11; Weingast & Moran, 1983). Consequently, a seven category issue coding scheme was created. The categories were defined as follows:

(A) Content: policy decisions involving the regulation of broadcast content (examples: political programming, indecency).

(B) Structural: policy decisions involving the regulation of ownership of broadcast stations or networks (examples: cross- and group-ownership regulations; attributable ownership specifications).

(C) Technical: policy decisions pertaining to technological requirements, availability, or specifications (examples: antenna and transmitter issues).

(D) Procedural: policy decisions pertaining to FCC and broadcaster operational procedures (examples: reporting requirements, program log requirements).

(E) Licensing: policy decisions pertaining to the criteria for broadcast station license applications, renewal, and retention (examples: applicant criteria, renewal requirements).

(F) Social: policy decisions pertaining to broader social issues as they relate to broadcast operations (examples: EEO requirements, environmental impact regulations).

(G) Other: policy decisions not fitting one of the above criteria or simultaneously meeting two or more of the above criteria.

Intercoder reliability for this variable was .92 using Scott's pi statistic.

Stakeholder Comments

Stakeholder comments were obtained from the Record Image Processing System (RIPS) available in the Public Reference Room of the Federal Communications Commission. Comments were analyzed during a 3.5 month period during the summer of 1998. RIPS contains scanned images of all formal comments filed in regards to FCC decisions, dating back to the beginning of 1992. Until recently, RIPS was the only electronic method by which members of the public could have ready access to all of the formal comments filed in regards to FCC decisions. In the past couple of months, RIPS has been replaced by the Electronic Comment Filing System (ECFS). ECFS is accessible via the FCC's home page on the World Wide Web (www.fcc.gov). Consequently, analyzing stakeholder comment data longer requires spending hours at the RIPS terminal at the Commission's headquarters in Washington, D.C. In all, well over 1,000 sets of stakeholder comments were analyzed.

The stakeholder comments were content analyzed according to two primary criteria. First, the comments were content analyzed in terms of whether they favored or opposed the policy proposal presented by the FCC. Intercoder reliability for this variable was .95 using Scott's pi. Of course, not every commenter addressed every policy issue contained within a decision document. Some commenters systematically addressed each individual policy issue, while, at the other extreme, some commenters addressed none of the individual issues, instead addressing other issues entirely, or advocating caution or timeliness in general (such comments were excluded). For the individual policy issues, the number of comments filed ranged from zero to forty-four.

The identify of the commenter was recorded as one of 13 possible stakeholder categories. These categories were created by examining a sample of the comments and identifying recurring commenter types. These categories and their associated descriptions are outlined in Table One. Intercoder reliability for this coding scheme was .98 using Scott's pi. ⁽⁴⁾ These 13 categories were eventually collapsed to reflect the broader stakeholder categories identified in the research questions. Specifically, broadcast industry associations, the NAB, auxiliary broadcast organizations, commercial broadcasters, and non-commercial broadcasters, were all collapsed into a general broadcast industry category. The competing industry and individual categories remained independent. The remaining categories (electronics industry, production industry, research institutes, government organizations, other, and public interest groups) were collapsed into a final "all others" (REST) category.

This approach to collapsing the variables was done for a number of reasons. First, from a theory testing and research standpoint, fairly broad categories of stakeholders (e.g., industry, the public, etc.) have typically been the focus of analysis in the research on regulatory behavior. The research questions outlined above focused on the broad stakeholder categories most frequently identified in the literature on FCC behavior (with the above-noted exclusion of political stakeholders such as Congress and the White House) and these categories are reflected in the variable collapsing that was conducted here.

Second, the fairly limited number of cases (N = 89) presented a situation in which accounting for each stakeholder category separately in multivariate analyses would create too many independent variables in relation to the number of cases. Given that, for each stakeholder category, the number of comments favoring and the number of comments opposed to the policy proposal are included as separate independent variables, the number of independent variables would rise to 26 if all 13 stakeholder categories were included separately. This ratio of independent variables to cases is much lower than what is generally advised for reliable statistical analysis (Menard, 1995).

Finally, for many of the stakeholder categories, the frequency of participation was far too low to warrant including them as a separate independent variable. This situation is evident in Table Two, which presents the frequency of participation for each of the 13 stakeholder categories described in Table One. The table lists the number of policy issues, out of the 89 studied, in which the stakeholder group filed at least one comment addressing the issue. As the table indicates, commercial broadcasters commented on the highest number of policy issues (61 of the 89) followed closely by auxiliary broadcast organizations (56 of the 89) and the NAB (45 of the 89). Very low levels of participation were found among the production industry, research institutes, and public interest groups (3 out of 89 for each of these stakeholder categories).

The finding regarding the low level of participation among public interest groups is particularly striking and important. These results suggest that public interest group monitoring of -- and participation in -- the broadcast policymaking process is quite infrequent. Recall also that Research Question Number Three addressed the possible relationship between public interest group participation and decision outcomes. However, from a data analysis standpoint, public interest group participation does not occur frequently enough to warrant including the comments of this stakeholder group as a separate independent variable.

In the end, the policy decision and stakeholder comments data sets were combined in such a way that, for any individual policy issue, it was possible to identify the number of comments from each category of stakeholder both in favor of and opposing the policy proposal. The complete set of variables created for this analysis is summarized in Table Three.

Analysis

Logistic regression analysis was used to determine whether the independent variables helped predict whether or not a policy proposal was adopted. ⁽⁵⁾ Logistic regression does not produce an R² that is identical to the R² in linear regression, but a number of analogues have been proposed (Hagle & Mitchell, 1992; Menard, 1995). Of these "pseudo-R² s," the Aldrich and Nelson (1984, 1986) pseudo-R² performed best in a comparative analysis of four pseudo-R² s by Hagle and Mitchell (1992), thus it was employed (along with the adjustment recommended by Hagle and Mitchell) in the analyses presented here.

Logistic regression analysis also produces a classification table, which provides information on the number and percentage of cases correctly classified. However, given space limitations, these results will not be discussed here. The emphasis will instead be placed on the pseudo-R², following Menard's (1995) assertion that, within the context of theory testing, the R² is the more important measure of predictive power than case classification (p. 36).

Results

The overall distribution of decisions according to their dependent variable classification is presented in Table Four. As the table indicates, in 65 of the 89 policy issues (73 percent of the total), the policy proposal was adopted. The policy proposal was not adopted in 24 of the 89 cases (27 percent of the total). This breakdown indicates that policy proposals have a significant likelihood of being adopted once they are presented (² = 18.88, p < .01). In terms of the distribution of decisions across issue types, the Technical and Procedural (Broadcaster and FCC) categories accounted for almost 90 percent of the decisions, while the remaining issue type categories (Content, Structural, Licensing) accounted for just over ten percent of the decisions. These results are similar to the distribution of broadcast policy issue types found in Napoli's (1998b) analysis of the 1966 through 1995 time period. As Napoli (1998b) noted in that study, the more high profile issue types, such as content and ownership issues, comprise a very small portion of the FCC's broadcast policy decision output, yet these areas tend to receive a disproportionate amount of the attention from FCC behavior researchers.

Table Five outlines the mean levels of participation for each of the broad stakeholder categories. As the bottom right corner of Table Five indicates, the mean number of comments filed per issue was 9.19. It is important to recognize that the overall number of comments filed in any docket is typically much higher; however, when the docket is broken down into each of the individual policy proposals the Commission is considering, the average number of relevant comments decreases, given that not all commenters address all of the individual policy proposals presented by the Commission. As the far right column of this table indicates, content issues received the largest mean number of comments (17.33), followed by broadcaster procedural issues (11.73), technical issues (8.88), and FCC procedural issues (8.76). Much lower levels of participation were found for licensing (4.33) and structural (4.40) issues. In terms of participation across the stakeholder categories, broadcast industry participants not surprisingly demonstrated the highest levels of participation (5.78 per issue), followed by competing industry segments (1.55). Individual participation averaged less than one (.44) comment per policy issue, while the combined remaining stakeholder categories submitted an average of 1.43 comments per policy issue. Thus, the broadcast industry and competing industry segments were the most active monitors of the FCC's broadcast policymaking during the time period studied. It is also important to emphasize the fact that, in many instances, certain stakeholder categories submitted no formal comments regarding the individual policy issues, which brings down the overall mean for each stakeholder category.

Table 5

Mean Number of Comments by Issue Type

Table Six presents a correlation matrix of all of the independent variables and the dependent variable. The generally modest intercorrelations among the independent variables, along with the values of the tolerance statistics (not presented), indicate no significant multicollinearity problem. In terms of relationships with the dependent variable (DECISION), the bivariate correlations with BIFAV and COMPOPP are significant in what could be considered the expected directions. Specifically, there is a significant positive correlation (r = .29, p < .01) between DECISION and BIFAV, suggesting that the greater the number of broadcast industry comments in favor of a policy proposal, the greater the likelihood of that proposal being adopted. The significant negative relationship (r = -.33, p < .01) between COMPOPP and DECISION suggests that, as the number of competing industry comments opposing a policy proposal increases, the likelihood of the proposal being adopted decreases. It is interesting to note that there is no significant relationship between the dependent variable and the total number of comments in favor of (r = -.08, p > .05) and opposed to (r = -.18, p > .05) the policy proposal. These results suggest that the overall position and volume of comments filed has little influence on the FCC's broadcast policy decision outcomes. This possibility is further explored in Tables Seven and Eight.

Table 6

Correlation Matrix (N = 89)

Table Seven presents a cross tabulation between the decision outcome (policy adopted/not adopted) and the majority position of the entire body of commenters (majority in favor/majority opposed). The overall N declines because those instances in which there was no majority (e.g., equal number in favor and opposed), or in which no comments from any stakeholder group were filed, were dropped from the analysis. As the table indicates, in those instances in which a policy proposal was not adopted, the majority opposed the proposal seven times and favored it 12 times. In those instances in which the policy was adopted, the majority opposed the proposal 12 times and favored it 50 times. In this latter instance, the results strongly suggest a relationship between majority preferences and the decision outcome; however, the overall distribution is not significant at the .05 level (² = 2.48, p > .05).

Table 7

Decision Outcome by Majority Position (N = 81)

Table Eight presents the results of a logistic regression analysis in which the total number of comments in favor of the policy proposal (TOTALFAV) and the total number of comments opposed to the policy proposal (TOTALOPP) were entered as the sole independent variables. It should be noted that the regression coefficients presented in this and the following table are Bs, as opposed to standardized betas, as SPSS 8.0 does not calculate standardized betas. However, given that all of the independent variables are measured in the same units (number of comments), the Bs can be directly compared. This table (along with Table Nine) also contains information on the standard errors for the coefficients, the -2 Log LR (the statistic used to assess statistical significance for the independent variables) and the odds ratio associated with each coefficient (Exp(B)). An odds ratio greater than one indicates that the likelihood of a decision being adopted increases as the independent variable increases. An odds ratio of less than one indicates that the likelihood of a decision being adopted decreases as the independent variable increases. As Table Eight indicates, neither of the independent variables are significant at the .05 level, and the overall explanatory power of the model is quite low, though it is significant at the .05 level (adjusted pseudo-R² = .11, p < .05). Together, the results of Tables Seven and Eight suggest that simple majority preferences do not factor very strongly into the FCC's broadcast policy decision making.

Table 8

Logistic Regression Analysis of the Relationship Between Total Number of Comments and Decision Outcome (N = 89)

Table Nine presents the results of the logistic regression model with the collapsed stakeholder categories included as the independent variables. The results differ markedly from the results of the analysis presented in Table Eight. Both BIFAV (B = .30, p < .05) and COMPOPP (B = -.54, p < .05) are significant at the .05 level, suggesting that as the number of broadcast industry comments favoring a policy proposal increases, the likelihood of the proposal being adopted increases; and as the number of competing industry comments opposing the policy proposal increases, the likelihood of the proposal being adopted decreases. None of the other independent variables was significant at the .05 level. The overall explanatory power of the model is much greater than when all stakeholder comments were collapsed into favoring and opposing categories. The adjusted pseudo-R² for the model is .43 (p < .01), suggesting that the monitoring behavior of stakeholder groups, as measured by the volume and position of their formal comments, provides reasonably substantial and statistically significant explanatory power in terms of variance in the FCC's broadcast policy decision outcomes.

Table 9

Logistic Regression Analysis of Relationship Between Stakeholder Comments and Decision Outcome (N = 89)

Conclusion

This section first reviews the five research questions stated at the outset and offers answers to these questions on the basis of the results presented here. Next, this section offers some other general observations, drawn from the data analysis, about the dynamics of the broadcast policymaking process. Finally, this section discusses the limitations of the research design used here and offers some suggestions for future research.

Research Questions

Research Question #1: Is there a relationship between the overall distribution of publicly filed comments and the FCC's broadcast policy decision making?

The results presented here indicate that the overall distribution of formal comments has no relationship to the FCC's broadcast policy decision outcomes. Whether the majority of commenters approves or disapproves of a particular policy proposal appears to have no bearing on whether or not the Commission decides to adopt the proposal.

Research Question #2: Is there a relationship between the quantity and position of comments filed by the general public and the FCC's broadcast policy decision making?

The results presented here indicate no significant relationship between the quantity and position of comments filed by the general public and the FCC's broadcast policy decision outcomes. Representatives of the general public were not particularly active in monitoring the Commission's broadcast policymaking activity. Individuals filed comments in only 17 of the 89 issue studied, and the overall volume of comments filed in those 17 instances was relatively low (mean of 2.29, maximum of two). Thus, it is perhaps not surprising that the participation of individuals bore little relationship to the Commission's decision outcomes. In any case, the results here confirm the frequently made observation that the public generally pays very little attention to the activities of the FCC and has little effect on its behavior.

Research Question #3: Is there a relationship between the quantity and position of comments filed by public interest groups and the FCC's broadcast policy decision making?

Given the very low levels of participation of public interest groups in the sample of broadcast policy decision studied here (public interest groups filed comments in only three of the 89 decisions studied), it was impossible to meaningfully investigate the potential influence of public interest groups on the FCC's decision making. However, the low level of participation documented here suggests that public interest groups are not particularly active monitors of the FCC's behavior and are thus not likely to be capable of exerting much influence over the Commission's decision making.

Research Question #4: Is there a relationship between the quantity and position of comments filed by the broadcast industry and the FCC's broadcast policy decision making?

The monitoring activity of the broadcast industry does appear to be significantly related to the FCC's broadcast policy decision outcomes. Specifically, the results presented here suggest that the intensity of broadcast industry support for a policy proposal is significantly related to the likelihood of that proposal being adopted. Of course, the broadcast industry was by far the most active participant/monitor of the Commission's decision making of all of the stakeholder groups studied here. This participation appears to translate into influence over decision outcomes. Thus, as far as the broadcast industry is concerned, participation in the decision making process is important for maximizing the likelihood that the decision outcome reflects the broadcast industry's policy preferences.

Research Question #5: Is there a relationship between the quantity and position of comments filed by competing industry segments and the FCC's broadcast policy decision making?

The results presented here indicate that competing industry segments stand alongside the broadcast industry as an important source of influence over the FCC's broadcast policy decision making. This is perhaps not surprising given that competing industry segments represented the second most active monitor of the FCC's broadcast policymaking activity. The results indicated that the number of comments opposing a broadcast policy proposal were significantly related to the likelihood of that policy not being adopted. These results suggest that, when competing industry segments view a broadcast policy proposal as threatening to their own interests, they are capable of influencing the Commission to reject such proposals. These results illustrate the common critique that is leveled at the traditional "industry capture" theory of FCC behavior. Specifically, such theoretical approaches need to become more sophisticated and account for the fact that, even within a relatively narrow component of the FCC's policymaking (in this case, broadcast policymaking), there are multiple interested industry segments with separate and distinct policy preferences, each of which is attempting to influence the Commission to act in accordance with its preferences. Thus, to say that any one industry segment has "captured" the FCC vastly oversimplifies the dynamics -- and the reality -- of the policymaking process.

General Observations

The conclusions drawn thus far need to be tempered by the fact that the overall distribution of decision outcomes indicates that, once a policy proposal is forwarded, it is very likely to be adopted. It is important to recognize that the analysis presented here focused on only the latter stages in the policymaking process (comments filed in response to NPRMs). It is possible that a better understanding of the dynamics of stakeholder influence could come from investigating earlier stages in the process -- specifically, the time period in which policy options are first being generated. If stakeholder influence extends to this earlier stage in the policy process, then the results presented here may underestimate the degree to which stakeholder groups are capable of affecting policymaking. In any case, future research should focus on the formulation of policy proposals.

In terms of the intensity of stakeholder monitoring, the results presented here suggest that monitoring varies significantly by issue type. It was interesting to note that content issues received by far the greatest volume of comments, particularly from the broadcast industry (the content category also produced the highest mean levels of participation from individuals and the REST category as well), suggesting that any potential regulation of content attracts the greatest levels of interest and participation from stakeholder groups. Of course, this conclusion must be tempered by the fact that very few content issues (a total of three) were contained within the data set.

Finally, the fact that the broadcast industry and competing industry stakeholder categories were the only ones whose comments were significantly related to the FCC's decision outcomes suggests that only those groups that are the most active in monitoring the behavior of the FCC and participating in its decision making process are capable of influencing the ultimate decision outcomes. Thus, intensity of monitoring does appear to matter. Those stakeholder groups that monitor less actively have much less influence potential.

Research Issues

This study represents the first effort to incorporate a quantitative analysis of stakeholder comments into a model of FCC behavior. Of course, the FCC's behavior is no doubt a function of a number of other factors, including political influence from Congress and the White House (see Napoli, 1998a, 1998b) and its own internal values and interests. These factors were not included in the model presented here and should be incorporated into future research that attempts to apply the agency theory framework to the behavior of the FCC. Despite this limitation, this study has provided evidence that the intensity of monitoring can affect the FCC's decision outcomes.

A larger sample of decisions would facilitate more sophisticated analyses than those presented here. For instance, the possibility of stakeholder influence varying by issue type could be better explored. It would also be possible to focus on narrow categories of stakeholder groups, or to focus on only those cases in which certain stakeholder groups actually did participate (e.g., analyze only the subset of decisions in which individuals or competing industry groups participated). Indeed, it is important to recognize that is impossible to draw firm conclusions about the relative influence power of various stakeholder groups unless the intensity of their participation in the process is held constant. However, the results presented here do suggest that influence is, to a certain degree, a function of participation, which is an important conclusion for those hoping to influence the FCC's decision outcomes.

Endnotes

1. For an extensive review of this literature, see Napoli (forthcoming).

2. An interview with the head of the Library Branch of the FCC's Office of Public Affairs confirmed that the LEXIS database was likely to provide a more thorough record of FCC decisions than print sources such as the FCC Record or the Federal Register (Thomas, 1997).

3. It was necessary to analyze only a sample of the 34 decisions due to the enormous volume of comments (both in terms of numbers and page length) filed for any one decision and the limited time period in which access to the Record Image Processing System at the Federal Communications Commission was possible.

4. Comments for 20 of the 92 decision issues (22 percent of the total) were recoded by the second coder. Within these 20 decision issues, a total of 201 comments were recoded.

5. Logistic regression is a variation of standard linear regression that is designed specifically for dichotomous dependent variables (see Hosmer & Lemeshow, 1989; Menard, 1995). Logistic regression is preferable to OLS regression for dichotomous dependent variables for a number of reasons. Unlike linear regression, the logistic regression model always yields predicted probabilities between 0 and 1. Also, logistic regression overcomes the heteroscedasticity inherent in using a linear function for a binary dependent variable (Demaris, 1995, p. 957; Morgan & Teachman, 1988, p. 933).

References

Alchian, A.A. & Demsetz, H. (1972). Production costs, information costs, and economic organization. American Economic Review, 62, 777-795.

Aldrich, J.H. & Nelson, F.D. (1984). Linear probability, logit, and probit models. Beverly Hills, CA: Sage.

Aldrich, J.H. & Nelson, F.D. (1986). Logit and probit models for multivariate analysis with qualitative dependent variables." In W.D. Berry & M.S. Lewis-Beck (Eds.), New tools for social scientists: Advances and applications in research methods Beverly Hills, CA: Sage.

Berner, R.O. (1976). Constraints on the regulatory process: A case study of regulation of cable television. Cambridge, MA: Ballinger.

Braun, M.J. (1994). AM stereo and the FCC: Case study of a marketplace shibboleth. Norwood, NJ: Ablex.

Canon, B.C. (1969). Voting behavior on the FCC. Midwest Journal of Political Science, 13, 587-612.

Cohen, J.E. (1986). The dynamics of the "revolving door" on the FCC. American Journal of Political Science, 30(4), 680-708.

Cole, B. & Oettinger, M. (1978). Reluctant regulators: The FCC and the broadcast audience. Reading, MA: Addison-Wesley.

Crotts, G. & Mead, L.M. (1979). The FCC as an institution. In L. Lewin (Ed.), Telecommunications: An interdisciplinary study (pp. 39-119). Dedham, MA: Artech House.

Demaris, A. (1993). Logit modeling: Practical applications. Thousand Oaks, CA: Sage.

Edelman, M. (1950). The licensing of radio services in the United States, 1927-1947: A study in administrative formulation of policy. Urbana: University of Illinois Press.

Federal Communications Commission (1997). In the matter of amendments of Parts 73 and 74 of the Commission's rules to permit certain minor changes in broadcast facilities without a construction permit. 12 FCC Rcd 12371.

Gormley, W.T. (1979). A test of the revolving door hypothesis at the FCC. American Journal of Political Science, 23(4), 665-683.

Hagle, T.M. & Mitchell II, G.E. (1992). Goodness-of-fit measures for probit and logit. American Journal of Political Science, 36(3), 762-784.

Hosmer, D.W., Jr. & Lemeshow, S. (1989). Applied logistic regression. New York: Wiley.

Jensen, M.C. & Meckling, W.H. (1976). Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3, 305-360.

Krasnow, E.G., Longley, L.D. & Terry, H.A. (1982). The politics of broadcast regulation (3rd ed.). New York: St. Martin's Press.

LeDuc, D.R. (1973). Cable television and the FCC: A crisis in media control. Philadelphia: Temple University Press.

Levine, M.E. & Forrence, J.L. (1990). Regulatory capture, public interest, and the public agenda: Toward a synthesis. Journal of Law, Economics, & Organization, 6, 167-198.

Lichty, L.W. (1962). The impact of the FRC and FCC commissioners' background on the regulation of broadcasting. Journal of Broadcasting, 6, 97-110.

Linker, J. (1983). Public intervenors and the public airwaves: The effect of interest groups on FCC decisions. In J.J. Havick (Ed.), Communications policy and the political process (pp. 149-170). Westport, CT: Greenwood Press.

McGregor, M.A. (1986). The FCC's use of informal comments in rule-making proceedings. Journal of Broadcasting & Electronic Media, 30(4), 413-425.

McQuail, D. (1992). Media performance: Mass communication and the public interest. Newbury Park, CA: Sage.

Menard, S. (1995). Applied logistic regression analysis. Thousand Oaks, CA: Sage.

Morgan, S.P. & Teachman, J.D. (1988). Logistic regression: Description, examples, and comparisons. Journal of Marriage and Family, 50(4), 929-936.

Napoli, P.M. (1997). Regulatory behavior and the Federal Communications Commission: An analysis of broadcast policy making and enforcement activity. Unpublished doctoral dissertation, Northwestern University.

Napoli, P.M. (1998a). Government assessment of FCC performance: Recurring patterns and implications for recent reform efforts. Telecommunications Policy, 22(4/5), 409-418.

Napoli, P.M. (1998b, November). The FCC and broadcast policy decision making: A test of multiple theories of regulatory behavior. Paper presented at the annual meeting of the National Communication Association, New York, NY.

Napoli, P.M. (forthcoming). FCC behavior: Theoretical perspectives and considerations for future research. To be published in Communication Yearbook.

Schwartz, B. (1959). The professor and the commissions. New York: Alfred A. Knopf.

Stigler, G.J. (1971). The theory of economic regulation. Bell Journal of Economics and Management, 2(2), 3-21.

Symons, H.J. (1989). The communications policy process. In P.R. Newberg (Ed.), New directions in telecommunications policy, vol. 1: Regulatory policy: Telephony and mass media (276-300). Durham, NC: Duke University Press.

Teske, P.E. (1990). After divestiture: The political economy of state telecommunications regulation. Albany, NY: State University of New York Press.

Thomas, G.J. (1997, June 27). Personal interview.

Weingast, B.R. (1989). The congress-bureaucratic system: A principal agent perspective (with applications to the SEC). Public Choice, 44(1), 147-191.

Weingast, B.R. & Moran, M.J. (1983). Bureaucratic discretion or congressional control? Regulatory policymaking by the Federal Trade Commission. Journal of Political Economy, 91(5), 765-800.

Williams, R.J. (1976). The politics of American broadcasting: Public purposes and private interests. Journal of American Studies, 10(3), 329-340.

Woolley, J.T. (1993). Conflict among regulators and the hypothesis of congressional dominance. Journal of Politics, 55(1), 92-114.