Future Expansion of the NANP

updated July 9, 2001

Epilogue: (February 21, 2021) Clearly, the NANP did not in fact run out of 10-digit telephone numbers by 2007, and current projections have pushed that date well past 2035. This proposal is left as it appeared here in 2001; there will be plenty of time to revisit the details before any expansion plan is needed.

Beginning in 1995, telephone numbers in North America are of the format NXX-NXX-XXXX. Each N represents a digit between 2 and 9, and each X represents any digit. The first three digits are the Numbering Plan Area code or NPA (“area code”), the next three digits the Central Switching Office Designation or CSOD (“prefix”), and the last four digits the Subscriber Line Identifier or SLID. Each NPA can contain almost 800 CSODs; the number is slightly less than 800, because certain combinations such as 911, 976, 555, and 950 are reserved for special purposes, and other combinations such as the numerics of the NPA or an adjacent NPA are generally avoided if possible. Thus, the total capacity of each NPA is slightly less than 8 million numbers. In practice, however, the capacity of each NPA is much lower, because each CSOD must be assigned to a specific geographic territory. Unused capacity from one CSOD cannot easily be “borrowed” by a central office whose CSOD(s) are full. (It is possible under some circumstances to subdivide a CSOD by the “thousands” digit, but this is seldom done.) There are theoretically 800 possible NPAs, but again, certain combinations are not used. All combinations with 9 as the middle digit are reserved for future expansion of the numbering plan. All combinations with the second and third digits the same (e.g., 233, 577, 622, etc.) are reserved as “easily recognized” codes for non-geographic services such as existing 500, 800, 888, and 900 services. Certain other combinations are reserved, leaving about 600 possible area codes.

The NANP is currently experiencing a period of feverish expansion in the number of area codes. There are several factors that caused a temporary spike in activity — pent-up demand that was held off until the new area codes were made available in 1995, political concerns that can now be addressed (assigning each Caribbean territory its own separate area code, and giving a separate area code to the northern territories of Canada), and the assignment of prefixes to prospective competitors for local service. However, there are other factors fueling demand that will continue unabated. Current estimates indicate that some time around the year 2025 to 2035, or possibly as early as 2005 to 2007, the current system of ten-digit telephone numbers will reach its capacity, and we will then need to expand to longer numbers.

Several proposals are being discussed for the specifics of longer numbers. It is important to reach a decision many years before the plan is put into place, in order that the necessary preparations can be performed as part of the routine maintenance on the telephone network. We can’t change the numbering plan architecture overnight; for discussion of some of the issues involved, see “Why Not 8-digit Phone Numbers?” on this web site.

It is generally expected that within the next few years, all calls within the NANP will be dialed with the entire 10-digit number, even if the NPA is the same and the call is local. This raises the possibility that the restriction that the first digit of the CSOD not be 0 or 1 might be lifted. However, CSODs beginning with 0 or 1 are currently used for certain special billing purposes and non-dialable numbers. A new scheme could be devised for those special billing numbers, but it would provide only a 25% increase in capacity within each NPA, which would provide a reprieve of only a scant few years in high-growth metropolitan areas. Further, consumers are proving stubborn in accepting mandatory 10-digit dialing.

Most proposals currently circulating envision either an 11- or 12-digit North American number: a 4-digit area code, plus a 7- or 8-digit local number. The area code will be of the format NXXX, although there may be a range reserved for yet another future expansion. (If we can’t make do with 8,000 area codes for the NANP, though, we’re in serious trouble.) The local number may remain in the form NXX-XXXX or NXXX-XXXX, or it may be generalized to any arbitrary 7- or 8-digit number; i.e., XXX(X)-XXXX.

The most commonly proposed method of expanding to longer-than-ten-digit numbers is to use the N9X range of area codes to expand the area code to four digits. Under this plan, each existing area code would have a 9 inserted between the first and second digits: 415 would become 4915, and so forth. This would allow permissive dialing, since there would be no area code 491 to create ambiguity with area code 4915. Eventually, the old three-digit area codes would be phased out, allowing arbitrary four-digit area codes to be used. This plan alone would create several thousand new area codes and add a factor of ten to the total capacity of the NANP.

The Industry Proposal

The Industry Numbering Committee (INC) has been developing recommendations for the expansion of the NANP numbering format beyond 10 digits. Unfortunately, the INC has reached consensus on a number of faulty assumptions, leading their proposal down an untenable path.

The INC assumes, first of all, that, before the introduction of longer numbers, we will first dial all calls within the NANP as 10 digits, without regard to whether the NPA is the same or whether or not the call incurs a toll charge. The use of the '1+' prefix must be completely eliminated, even as an optional feature. As a replacement for the current practice of “toll alerting” (requiring all toll calls to be dialed as 1+10D, as is currently the case in about 2/3 of the U.S. and all of Canada), the INC proposes that telcos offer an optional feature that any toll call dialed would be greeted with a brief recorded announcement or special tone. It is extremely unlikely that such a proposal would be acceptable to the PUCs of every state that currently has “one-plus” toll alerting. The INC’s plan is explicitly contingent upon universal adoption of its unrealistic Uniform Dialing Plan — a condition which will likely never be met (at least so long as an international call costs noticeably more than a local call).

The INC currently advances a plan to add either 0 or 1 after the existing NPA, and possibly also adding the same digit on the front of the existing local number. Their draft proposal incorporates a request to give Canada a separately identifiable numbering space by reserving one of those digits for Canada’s permanent exclusive use. Thus, for instance, 202-555-0123 would become either 2021-555-0123 or 2021-1555-0123, while 613-555-0174 would become either 6130-555-0174 or 6130-0555-0174. There would be a permissive dialing period during which it would be possible to dial either 10D or the new 11D or 12D number. After permissive dialing ends, the number format would be expanded to allow any arbitrary NXXX area code and any arbitrary 7- or 8-digit local number. The 7-digit number space would be increased by 25% due to the use of all ten digits in the first position; the 8-digit numbers would offer a 12-fold increase per area code.

My New Proposal

The pace of adding new area codes is only just beginning to slow with the belated introduction of number conservation measures such as thousands-block pooling and rate-center consolidation, raising the spectre of a premature exhaust of the entire 3-3-4 number format for North American numbers. It is thus prudent to examine means by which the expansion to longer-than-10-digit numbers can be forestalled. My original plan is still available on this site, but I have recently made some changes.

All area codes with the N9X format (that is, all area codes with ‘9’ as the middle digit) have been reserved for expansion of the NANP number format. Releasing those codes for assignment would provide a substantial boost to the supply of usable numbers, of similar magnitude to proposals to allow area codes and/or CSODs to begin with ‘1’ and/or ‘0’. The catch, of course, is that releasing those codes for present use could complicate the eventual migration to 11- or 12-digit numbers.

I still propose a migration directly from 10- to 12-digit numbers, but with some modifications. First, rather than the “9 + 3 = 12” mnemonic, I propose to use the mnemonic “1 + 2 = 12,” in order to free up 72 additional 3-digit area codes. Each current 10-digit number NPA-NXX-XXXX would become, in the new scheme, NPA1-2NXX-XXXX. The additional digit on the area code portion would be added at the end, rather than in the second position. There would be no confusion with existing 10-digit numbers during the transition, because NPA-12N-XXXX is not a valid user-dialable number under existing rules.

However, there are some non-user-dialable numbers of the format NXX-[0/1]XX-XXXX, used for billing and network routing purposes, so there would be a potential conflict. To resolve this conflict, I propose a modest set of measures. First, freeze all new assignments of numbers of the format NXX-01X-XXXX, NXX-11X-XXXX, and NXX-12X-XXXX. Allow a period of time for attrition to minimize the cost and inconvenience of displacing any existing numbers in those ranges. Since the customers do not dial these numbers, though, the inconvenience would be minimal. The greatest direct effect on a customer would be a possible change in the number on a non-subscriber calling card. Then allow a period of time for gradual reassignment of any remaining numbers in the affected ranges. If a period of, say, two years is provided, the impact on the telcos will be minimal, with numbers changed over in small batches during lulls in the normal workload. Further, any burden imposed on the telcos by this procedure would be far less than that imposed by prematurely converting to longer numbers.

Once all existing non-user-dialable numbers in the affected ranges are reassigned, the NANP will be ready for the next step, which will be triggered when the supply of area codes dwindles below an agreed threshhold. Existing dialable numbers will convert by adding a ‘1’ to the area code and a ‘2’ to the front of the prefix. Thus, for instance, 628-555-xxxx will become 6281-2555-xxxx. Existing non-customer-dialable routing numbers based on actual area codes will also add a ‘1’ to the area code, but will prefix the remainder of the number with another ‘1’; thus, 628-075-xxxx will become 6281-1075-xxxx. Billing numbers based on RAO codes (of the same format as area codes, but a separate numbering space) will add a ‘0’ to the end of the RAO code and a ‘1’ to the front of the remaining digits; thus, 747-090-xxxx will become 7470-1090-xxxx. The earlier migration of the conflicting numbers will allow permissive use of the 10- and 12-digit routing and billing codes in parallel. After the migration is complete, each area code could open up assignment of local prefixes beginning with digits other than ‘2’: either NXXX or XXXX format.

This proposal would create area codes like 3611 in south Texas, which might create some confusion with N11 short codes. It would be appropriate to do human factors studies to determine the impact of such confusion and whether it would outweigh the advantages of the “1+2="12"” mnemonic. Other issues, such as a possible separate migration path for 950 numbers or 976 numbers, could also be examined.

As far as 99% of the public is concerned, however, the big change will be that their regular numbers will change from NXX-NXX-XXXX to NXX1-2NXX-XXXX, with a year or two for permissive dialing. Of course, any areas that still have 7-digit dialing will have to either convert to 10-digit dialing before the beginning of permissive 12-digit dialing, or else undergo a flash cut from 7d to 8d for local calls. While such a flash cut imposes a certain amount of disruption on both telcos and customers, the overall benefits of the scheme outweigh the costs.

The LincMad Proposal, mark 2:

Remember, 1+2='12'! 628[1]-[2]555-0174

This proposal would expand the numbering capacity of the North American Numbering Plan by a factor of 100, which should allow ample room to assign a telephone number to every person, computer, fax machine, pet, and garden gnome. A twelve-digit numbering space would allow more than 500 billion possible numbers. In addition, by freeing up the N9X range of three-digit area codes for assignment, an additional 500 million numbers would be made available before abandoning the 10-digit number format.

Some proposals add an eighth digit to the end of the local number, so that, for instance, 555-0123 becomes 555-01230. Adding the extra digit to the CSOD rather than at the end of the SLID has the advantage that it does not magnify the inefficiency in number utilisation caused by the requirement that each town have its own prefix, since there will now be nearly 8000 (or possibly 10,000) CSODs in each NPA, but each CSOD will still represent only 10,000 SLIDs. A further advantage is that the introduction of 12-digit numbers will usher in a lengthy period of stability in NPA assignments, since it will take years for even the fastest-growing NPA to absorb a tenfold increase in capacity.

A ten-fold increase in capacity might be adequate to last for decades or even centuries; then again, why not go for the 100-fold increase and be certain? In particular, the change from N[0/1]X to NXX area codes more than quadrupled the supply of numbers, but some projections show us squandering that new capacity in as little as 10 to 12 years — counting from 1995!

The major advantage of this plan over the INC’s current plan is that this plan does not rely on the assumption that all 50 states and all 18 foreign governments will adopt the INC’s unrealistic Uniform Dialing Plan. This plan can even be implemented with local 7D dialing migrating to 8D, although a flash cut would be required. However, such a flash cut is not without precedent. London, England, changed its numbers from +44 171 NXX XXXX to +44 20 7NXX XXXX, and similarly +44 181 NXX XXXX to +44 20 8NXX XXXX, in 2000. Although there was permissive dialing for the full national or international number, local dialing was flash cut from 7 to 8 digits. Disruption was minimal and short-lived. The flash cut was performed on a Sunday, and within a matter of hours the rate of incorrectly dialed calls receded to acceptable levels.

The INC’s Uniform Dialing Plan even contradicts itself. It gives as the ideal UDP a scheme in which all calls are dialed as 10D, whether local or toll, same or different NPA, and yet it also states “A uniform dialing plan does not preclude local jurisdictions from permitting non-conflicting alternatives (e.g., 7-digit local dialing).” However, 7D local dialing is fundamentally incompatible with uniform 10D dialing of all calls. The INC also calls for the universal elimination of the 1+ prefix without offering any concrete alternative method of toll alerting.

Likewise, my plan allows for phased implementation of the Release of the D Digit. The D digit is the first digit of the CSOD; in other words, the N in npa-Nxx-xxxx. It is currently restricted to the values 2 through 9, but allowing any arbitrary number in that slot would provide a 25% increase in capacity, irrespective of the other details of the numbering plan. Release of the D digit is conditioned on elimination of local dialing: the number 023-4555 cannot coexist with 0-234-555-0123. The INC thus requires that all parts of the NANP dial 10 digits, and only 10 digits, for all calls. My plan would allow for the retention of toll alerting by the 1+ prefix, and would also allow each area code or state/‌province/‌territory to decide independently to release the D digit.

This plan allows for several compatible dialing plans to coexist. An area that wanted to retain 8-digit local dialing could do so by preserving the restriction of the D digit. Thus, numbers in such an area would be in the form NXXX-NXXX-XXXX. An area that wanted to increase the capacity per area code could do so by releasing the D digit, mandating 12-digit dialing.

The toll alerting issue is separate from the issue of allowing some calls to be dialed without the area code. Ideally, all parts of the NANP would permit 1+ on all calls, but require it only on toll calls, but the reality is that we are unlikely to converge on a common dialing plan any time soon, and it is foolish to predicate NANP expansion on such a precondition.

There is a close analogy between this situation and the introduction of interchangeable CO codes in the 1970s, 80s, and 90s. Many areas of the United States used either 1+7D for toll calls within the same area code, or else straight 10D without the 1+ prefix for toll calls to other area codes. Those areas could keep their dialing plan by preserving the NNX-XXXX format of local numbers. However, as number crunches hit, first in Los Angeles and New York, and later in other parts of the country, individual regions could opt to generalize local numbers to NXX-XXXX, at the cost of eliminating 1+7D and 10D dialing, at least as a general default. (1+7D was eliminated entirely, and 10D dialing was restricted to at most a small regional calling area.) Much of the NANP was able to retain the N[0/1]X-NNX-XXXX number format until 1995, when area codes were generalized.

The plan presented here allows for a 100-fold increase, or potentially as much as a 125-fold increase, in the numbering capacity of the NANP, while at the same time allowing for local options such as toll alerting and abbreviated local dialing. It would cause substantially less disruption to end users, compared to the INC’s draft plan.