Does allowing the bells to offer InterLATA long-distance service affect entry into local telephony?

• Author: Flaherty. Susan M.V.

1. Introduction

   One of the primary goals of the watershed Telecommunications Act of 1996 is to remove the structure of legal monopoly in the U.S. local telephone exchange market held by the Regional Bell Operating Companies (RBOCs). (1) The Act requires the RBOCs to lease out or "unbundle" specific bottleneck components (commonly referred to as Unbundled Network Elements, or UNEs) of the local public-switched telephone network to Competitive Local Exchange Carriers (CLECs) to facilitate the latter's entry into the local telephone exchange market. (2) The Act allows CLECs to obtain unbundled transport, switching, and loop facilities collectively from the RBOCs (in what is known as a UNE-P arrangement) or install their own switching equipment while leasing RBOC transport/loop facilities. (3) Ultimately, the Act envisions a highly competitive local telephone exchange market served ubiquitously by both intramodal (e.g., self-sustained, facilities-based CLECs) and intermodal (e.g., cable, wireless, Internet) service providers who would compete directly against the RBOCs.

   Recognizing that the RBOCs have little incentive to unilaterally open their local exchanges to potential competitors, section 271 of the Act allows the RBOCs to enter the InterLATA (interexchange or long-distance) market (4) on a state-by-state basis conditional on their demonstration that the local exchange is open to competitors. (5) Prior to the Act, RBOCs were restricted from entering the InterLATA markets. The reason for this restriction was the fear that the RBOCs would leverage their monopoly positions in the local exchange market into the in-state interexchange access market (6) and foreclose competing long-distance providers (i.e., effectively reestablishing the jointly monopolized local and long-distance market structure that existed before 1984). The establishment of alternative local exchange providers as envisioned by the 1996 Act would mitigate such an outcome by creating a new competitive market for interexchange access services in addition to a competitive market for local access services.

   The purpose of this paper is to empirically evaluate whether the market-opening mandates of section 271 have been successful, that is, whether local entry by partial facilities-based CLECs has actually occurred in those states where the RBOC has been allowed to enter the long-distance market. The remainder of the paper proceeds as follows. Section 2 describes the data and empirical methodology employed to evaluate the impact of RBOC entry into long-distance on competitive entry into local telephony. Section 3 presents the estimation results and discusses the policy implications of section 271 from its "aggregate" impact across all authority states and, more appropriately, from its "disaggregate" impact on individual authority states. Results from the latter approach suggest that the section 271 experiment has been a success in the sense that it appears to exert a net positive influence on the rate of competitive entry into most state local exchange markets. Specifically, the results suggest that states where the RBOCs are granted the authority to provide in-region long-distance service will realize, on average, two additional CLECs serving their respective local service market relative to states that are not granted such authority. Section 4 discusses the findings in light of recent trends in the prices for local telecommunications services, while section 5 concludes the study.

2. Data and Empirical Methodology

   Market definition is a difficult problem that antitrust economists have spent a great deal of time studying and developing tools to assess. Ordinarily, a market is defined along at least two dimensions: product scope and geography. While the product dimension of a local telephone market is ever changing with technology characteristics not easily definable, the LATA remains the most appropriate definition of the relevant geographic market in which to assess the impact of CLEC entry given extant data constraints. For instance, Abel and Clements (2001, p. 232) note that when trying to explain market entry, the use of rule or market-based indicators is inappropriate; and other sources of data that ostensibly measure competitive entry are problematic. (7) For these reasons, this paper follows previous studies (e.g., Abel and Clements 2001; Zolnierek, Eisner, and Burton 2001; Abel 2002; Beard, Ford, and Koutsky 2002; Alexander and Feinberg 2004) and measures competitive entry into local telephony as a raw count of CLECs-deployed switches (i.e., by LATA).

   The CLEC counts employed in this analysis are derived from the Local Exchange Routing Guide (LERG) for the years 1996-2002 and correspond to CLECs who install their own switching equipment but lease transport and local loop facilities from the RBOCs as UNEs. Consideration of switch-based CLECs is an important issue for at least three reasons. First, these firms rely on unbundling elements provided by RBOCs to provide local telephone service. Since RBOCs must demonstrate that their local markets are open to competitors in part by provisioning these elements, an examination of switch-based CLECs allows for a direct test of RBOC entry into long-distance on competitive entry into the local exchange. Second, switch-based entry is frequently employed by RBOC competitors to provide local exchange service. Indeed, a report by the Association for Local Telecommunications Services (2003) classifies 48% of its listed CLEC entities as "primarily switched-based." In comparison, the next most common type, "rural ILEC-affiliated," constitutes only 25%. (8) Finally, while entrants have used UNE-P arrangements for serving the majority of their lines, the regulatory future of UNE-P has been far more uncertain relative to simple local loop unbundling. (9) Indeed, in March 2004, the D.C. Circuit Court vacated and remanded several of the Federal Communications Commission's unbundling rules, the most important of which mitigate the requirement that the RBOCs provision UNE-P. (10) As such, focusing the analysis on switch-based entry may be most appropriate for gauging the potential long-term impacts of the 1996 Act.

   Figure 1 illustrates the aggregate number of CLECs in the sample for each year of the panel. The graph shows a dramatic increase in the level of CLECs since the passage of the 1996 Act, with only a slight decrease occurring in 2001. (11) Table 1 presents the number of states (including Washington, D.C.) that had received section 271 authority by the end of each year and the proportion of the RBOCs' total end-user access lines constituted by these authority states. By the end of the sample period, 71% of all eligible states had received section 271 authority. (12) These states constitute approximately 75.1% of all RBOC end-user access lines.

   [FIGURE 1 OMITTED]

   It is assumed that CLEC entry by LATA is determined according to

   [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

   where the subscripts i, s, and t index LATAs, states, and years, respectively. The variable [alpha] denotes a constant term, while [epsilon] denotes the randomly distributed error. The variable [theta] denotes a vector of LATA dummies that are included in the model to control for any unobservable LATA-specific factors that are time invariant (i.e., fixed effects).

   The variable InterLATA is a vector of dummies pertaining to the timing of the authorization date of individual states' section 271 applications. Assume that a state receives section 271 authority during year t. The first dummy (Year of InterLATA Authority) indicates the year a state receives section 271 authority. The other dummies capture the pre- and postauthority impacts of a given state receiving section 271 authority, that is, the effects on CLEC entry over time. These variables include indicators for the first (First-Year Pre-InterLATA Authority), t - 1, and second (Second-Year Pre-InterLATA Authority), t - 2, years before a state received section 271 authority and indicators for the first (First-Year Post-InterLATA Authority), t + 1, and second (Second-Year Post-InterLATA Authority), t + 2, years following authority. Of course, to the extent that the mandates of section 271 are effective at facilitating the entry of CLECs, the coefficient estimates on each of these variables is expected to be positive.

   The variable [??] is a vector of covariates typically included in aggregate studies on competitive entry into local telephony. Since LATAs do not conform to any commonly recognized state or local government boundaries, data for covariates defined at the LATA level of aggregation do not exist. However, LATAs rarely cross state lines, and a wide range of potential explanatory variables that might influence CLEC entry are available at the state level of aggregation. As such, each of the covariates described here relates to the relevant state where a LATA is designated. (13)

   State regulatory regimes may affect the openness of an incumbent's local network to competitors. Therefore, once-lagged regulatory dummies are included in the analysis and are based on whether the state follows rate-of-return regulation (guaranteeing an RBOC a rate of return above its costs), has price cap regulation (where the incumbent RBOC operates under a price ceiling), or has a rate case moratorium (prohibiting an RBOC from requesting rate increases from the state regulatory authority) in effect. (14)

   The extent to which the incumbent is "entrenched" with heavy sunk costs in a market may indicate a credible commitment to drive out entrants with fierce price competition. The deployment of fiber-optic cables may represent such investments. Therefore, following Alexander and Feinberg (2004), the degree of entrenchment is controlled for by including the percentage of total kilometers of fiber deployed in each state controlled by the largest (in terms of total operating revenue)...