On December 15, the Federal Communications Commission is expected to vote on regulations requiring telephone companies to lease, or "unbundle," parts of their networks to competitors. If this triggers a sense of déjà vu, it should. The FCC has adopted rules on this subject three times since 1996-and all three times the rules have been struck down by the courts as overreaching.

On this fourth try, the FCC needs to get it right. This means limiting mandatory unbundling rules to markets where competition otherwise could not feasibly exist. Importantly, any such determination should take into account competition from new communications services, such as wireless and Internet telephony. The result would benefit consumers, as well as avoid an embarrassing fourth defeat for the FCC in court.

The rules under consideration stem from the Telecommunications Act of 1996, which required telephone companies to provide competitors with access to parts of their networks, such as switches and lines, at rates set by regulators. Exactly which parts were left to the FCC to decide. Under the chairmanship of Clinton-appointee Reed Hundt, the FCC drew up an extremely comprehensive list of such "unbundled network elements," or UNEs.

In 1999, however, the Supreme Court knocked down these rules as too far-reaching. The FCC then issued marginally revised rules. Once again the issue went to court, and once again, in 2002, the rules were struck down.

In 2003, the FCC issued a third set of rules, which eased requirements for advanced broadband services but kept other UNE requirements largely in place. These rules were adopted over the dissenting vote of Chairman Michael Powell, a Bush appointee. Predictably enough, this third set of rules was struck down, although they remain largely in effect under the FCC's transition plan.

The FCC's upcoming rulemaking-the agency's fourth bite at this particular apple-comes in response to this latest court ruling. Many, including some telephone companies who benefit from these regulations, have urged the Commission to tweak the rules marginally once again and keep them largely in place. If the rules were significantly scaled back, these companies argue, their costs would rise significantly, threatening their businesses.

The FCC should reject these arguments for several reasons. First, the Commission's goal should be a competitive marketplace, not the protection of individual competitors. There is simply no reason to subsidize favored companies in this market. Despite the image of telecom challengers as plucky start-ups, the field lobbying in favor of the thrice-rejected rules also includes some of the largest communications firms in the nation, such as MCI and AT&T.

But don't the rules help preserve competition and thus keep consumers' costs down? Not necessarily. In fact, the UNE rules can undercut real competition by encouraging potential challengers to lease existing telecommunications assets rather than construct their own facilities. The result is a Potemkin sort of competition among firms all using the same network, rather than real competition among independent providers each bringing network capacity to the table. Moreover, the sharing rules discourage incumbent telephone companies from investing in their existing networks, since they must share the benefit of any upgrades with their rivals.

In any case, the Commission's authority to impose broad access regulations is severely-and rightly-limited by the Telecommunications Act, as interpreted in the three court cases so far on unbundling. The cases turned on the language of the Telecommunications Act, which limits forced unbundling rules to situations in which access to the incumbent telephone company's network is "necessary" and the failure to have such access would "impair" a competitor's ability to offer services. According to these decisions, this language sets a high bar for the rules.

Thus, for legal and policy reasons, the Commission should substantially pare back these regulations so that forced access is required only where regulatory intervention is required for competition to survive. In other words, cases where the incumbent telephone company still holds-and will likely continue to hold-a "natural monopoly" on service.

Importantly, forced access should not be required just because a particular market does not have a competitor. The test should be whether a market is the kind that can support competition, not whether there is actually competition at any particular moment. This may sound odd, but it makes economic sense. Even if no competitors have yet entered a market, the potential for competition can still protect consumers. Conversely, FCC rules that mandate access to existing networks in such markets can themselves actually discourage such entry from happening.

At the same time, the FCC should not limit its analysis to any particular type of service or technology. Many of the strongest challenges to incumbents are not traditional, wireline-based telephone companies. Rather, the fastest growing options for consumers are coming from cable companies offering broadband services, Internet-based telephone services (which need not use telephone lines at all), and wireless phones. The FCC, in determining whether markets are subject to competition, should consider all such "intermodal" forms of competition.

The end result would be for most, if not all, elements to be freed from regulation, including switches and most lines in urban centers. Depending upon how intermodal competition is assessed, even the rules on local telephone loops to individual residences could be lifted.

The battle over these telecommunications competition rules has now gone on for eight years, moving from the FCC to the courts and back again three times. It is time for the FCC to end this regulatory soap opera by issuing narrow rules that can pass economic and legal muster. For the UNE rules, the fourth time can and should be the charm.

James L. Gattuso is Research Fellow in Regulatory Policy in the Thomas A. Roe Institute for Economic Policy Studies at the Heritage Foundation.