I. INTRODUCTION

 This purpose of this adjudication is to estimate the cost of providing 
telecommunications service in high-cost areas of this state.  Staff recommends 
that the Commission use the Hatfield Model HAI 5.0a, as run with the inputs and 
modifications made by Staff, to estimate the cost of service for high-cost 
areas served by US West, GTE, and Sprint.  Staff believes the HAI 5.0a model, 
as modified, provides the most accurate measurement of these costs.  Staff 
further recommends that the Commission use the embedded cost analysis which 
Staff has compiled using information provided by the rural companies, as the 
best estimate of the cost of service for the rural companies participating in 
this docket.
II. LEGAL PRINCIPLES

A. State Law
  1. Relationship between this docket and ESSB 6622 (Laws of 1998, chapter 337)

 The Commission is undertaking this investigation and adjudication to fulfill 
its responsibilities under the law, including ESSB 6622.  This 1998 legislation 
directed the Commission to provide estimates of the cost of providing 
telecommunications service in high-cost areas.  At a minimum, therefore, this 
proceeding needs to produce the information necessary to fulfill that 
particular legislative requirement.  Costs need to be estimated separately for 
one line per address and for all lines, and separately for residential and 
business customers. In addition to the requirements of section 1(2)(a) of ESSB 
6622, the Commission has required parties to provide the cost of a network 
serving only multiline subscribers.  See Guideline 2, Notice of Prehearing 
Conference, May 4, 1998.  
 However, the Commission has made it clear that this proceeding should be used 
to fulfill broader responsibilities than the transitory ones of a legislative 
study bill.  The preservation and advancement of universal service is an 
obligation that has been implicit in the Commission’s regulation since its 
beginnings, was made explicit by the 1985 Regulatory Flexibility Act, and was 
reinforced by the Telecommunications Act of 1996.  
 While the Commission initiated this proceeding before the ink was yet dry on 
ESSB 6622, the origins of this proceeding actually can be traced back to the 
Regulatory Flexibility Act. That law established a state policy in RCW 
80.36.300 with two potentially conflicting and contradictory objectives.  One 
was as old as AT&T itself:  to advance universal service.  A second was much 
newer: to promote competition.  Without competition, universal service was a 
natural outcome of the monopoly industry structure and the falling cost of 
telecommunications technology.  A monopolist wants everyone to have a phone, 
because a service that can reach everybody commands a higher value than one 
that doesn’t.  Falling costs made it easier to expand service into higher cost 
areas without imposing the burden of rate increases on existing customers.
 The introduction of consumer choice, or competition, as a policy objective 
required a fundamental change in approach for regulators.  While some aspects 
of competition could enhance universal service, for instance by driving down 
the cost of service for all customers, it also has the potential to undermine 
universal service.  The hallmark of competition is cost-based rates, and yet 
the essence of universal service is that rates are not to be excessive even 
when costs are.  With each move toward greater customer choice, the Commission 
has consistently heard the objections of incumbent companies that this 
competition will endanger universal service.  The concern is that competition 
will drive prices toward costs, thereby increasing rates in high-cost areas to 
the detriment of universal service.  That concern reflects the effort of the 
incumbent monopolists to protect their profits, but it also reflects a 
legitimate concern about the potential conflict between competition and 
universal service.
 Even without ESSB 6622, the Commission needs adjudicated determinations of the 
cost of providing telephone service in high-cost areas.  It needs this 
information so that it can advance the twin objectives of universal service and 
consumer choice.  Any given incumbent local exchange company’s revenues can be 
divided into a part that is necessary if it is to provide service in its 
high-cost areas and a part that can be attributed to its provision of service 
in other areas.  Those two parts require very different regulatory treatment.  
The part necessary for service in high-cost areas should be susceptible to 
competition only from other companies that are serving those high-cost areas; 
it should not be taken from a company that loses competitive business in 
low-cost areas.  The remaining part, which is not necessary to provide service 
in high-cost areas, should not be afforded such insulation from competition.  
The purpose of this proceeding is to draw that line -- to divide the revenues 
of companies serving high-cost areas into the part necessary to serve high-cost 
areas and the remainder that can be attributed to its service in other areas.
 The results of this case will be an input to the broader ongoing work of the 
Commission. Clearly, a primary use will be to fulfill the requirements of ESSB 
6622.  Beyond that, the Commission can use these results as a guide to the 
opening of local telephone markets.  The best example of this approach is the 
access charge reform rule adopted earlier this year.  The overall thrust of 
that reform is to make local companies compete for their access dollars by 
requiring that the bulk of the access revenue stream come from the 
more-competitive originating access service.  The Commission recognized that 
this pro-competitive policy had to be tempered, however, in recognition that 
some access revenues were used to maintain service in high-cost areas.  The 
rule therefore allows companies to collect that part of access revenues from 
the less-competitive terminating access service.  Companies will lose that 
revenue only if competitors go into the high-cost areas and offer service 
there.  This proceeding will provide the information required to implement the 
pro-competition and pro-universal service elements of the Commission’s access 
charge reform rule.
  2. Can or should the Commission take any action on universal service absent 
an explicit directive from the Legislature?

 This is a difficult question, because embedded in it are two erroneous 
assumptions.  One is that the Commission might “take action on universal 
service” in this proceeding.  The other is that there has been no “explicit 
directive” from the Legislature.
 The Commission will not in this adjudication take action on universal 
service.  The outcome of this adjudication will be a determination by the 
Commission of the cost of providing telephone service in high-cost areas.  Even 
the development of a state universal service program would not represent 
“taking action” because the product would simply be a recommendation to the 
Legislature, and in any event, the Commission’s decisions regarding that 
program will be made in the rulemaking phase and not in this adjudication.
 As discussed earlier, the “explicit directive” was given to the Commission by 
the Legislature in 1985 and by the Congress in 1996.  It certainly is true that 
the Commission does not have carte blanche to take any steps it deems 
appropriate to further universal service.  The state Supreme Court ruled that 
the Commission’s ratemaking authority does not extend to mandatory 
inter-company transfers of money, so a state program that parallels the federal 
program would require an “explicit directive” from the Legislature.  Beyond 
that limitation, the Commission can and should take steps to preserve and 
advance universal service.  Those steps will not be taken in this adjudication, 
but the Commission’s determination here will be most useful even without 
further “explicit directive” from the Legislature.
  3. Need for consistency between the universal service fund proceeding and the 
price of unbundled network elements. 

 The Commission need take no action in this proceeding to ensure consistency 
between the universal service fund and the prices of unbundled network 
elements.  Where such consistency is needed, it is best addressed in the 
development of the universal service program itself (which is a rulemaking and 
legislative issue and not part of this adjudication) and in the pricing of 
unbundled network elements (which is addressed in the second phase of Docket 
No. UT-960369).
B. Federal Law
 Congress has set out the basic requirements for the preservation and 
advancement of universal service, and the FCC has undertaken a great deal of 
work to implement these requirements.  The role of the FCC and the states is 
quite different with regard to universal service.  While Congress directed the 
FCC to do something about universal service, it did not require anything of the 
states.  Rather, in Sec. 254(f) Congress preserved the authority of states to 
take their own actions in furtherance of universal service.  Washington has 
signaled its intent to do so with the passage of ESSB 6622, building upon the 
earlier policy direction of the Regulatory Flexibility Act.
 Federal law, while not requiring state action, constrains state action.  
Congress anticipated that states might take actions that, while preserving 
universal service, would hinder competition, or that states might attempt to 
manipulate the federal system of government such that its burdens were shifted 
to other states or the federal jurisdiction.  These constraints must be kept in 
mind as the Commission engages in rulemaking on the universal service program 
and as it advises the Legislature on these issues.  It also must be aware of 
these constraints as it engages in ratemaking actions that are intended to 
preserve and advance universal service.  Examples include the WECA high-cost 
fund rate (the annual filing of which is under review in Docket UT-971140) and 
the access charge filings that will follow from adoption of the rule in Docket 
UT-970325.
 Congress did not, however, tell the states how to determine cost, and this 
adjudication is to determine cost and not to develop a universal service 
program.  The constraints of Sec. 254(f) therefore are not relevant to this 
proceeding.
III. POLICY ISSUES
A. Purpose of the Proceeding
 As discussed above, this adjudication was undertaken to provide one of the 
fundamental building blocks to a universal service support mechanism that is 
competitively and technologically neutral.  For it to be useful, it must be 
narrow.  The cost determinations can be used in the rulemaking phase, in 
various tariff filings, and in informing and advising the legislative process.  
Those other processes, however, must not be allowed to bleed into this narrow 
adjudication of the cost of providing telephone service in high-cost areas.  
While it is a fundamental building block, it is not the entire structure.
B. Timing/Implementation/Competition
 Some parties have argued that the Commission should take no action on 
universal service until competition has developed. Ex. 371T at 10 (Baker 
Responsive; AT&T); Ex. 387T at 19-20 (Johnson Rebuttal; Public Counsel).  Their 
reasoning is that as long as the incumbent is serving enough low-cost 
customers, relative to its high-cost customers, that its overall rates remain 
affordable, no explicit support mechanism is required.
 The Commission should consider that argument and reject it, but it should do 
so in the rulemaking proceeding and not here.  It should be rejected because it 
fails to appreciate that competition is not an alien force that will either 
arrive in our state or stay away.  The level and extent of competition is a 
direct function of the industry structure in this state.  If the state provides 
no explicit support for service in high-cost areas but instead relies on price 
averaging by incumbents, there will be virtually no competition in high-cost 
areas.  Explicit support is required if the Commission is to carry out its twin 
objectives of universal service and customer choice. Without explicit support, 
it will be forced to choose between those objectives.
 Having made this point, Staff again notes that the issue, while important, is 
not one that should be addressed in this adjudication.  There are crucial 
decisions that must be made about whether and when to implement a state 
universal service mechanism and how to coordinate such a mechanism with the 
federal counterpart.  Those decisions will require good information about the 
cost of providing telephone service in high-cost areas, and this proceeding can 
provide that information.
C. Selection of a Cost Model
 Staff has advocated in this proceeding that “the Commission use the exchange 
costs developed by Staff in this testimony to establish the amount of universal 
service funding,” and noted that, “The current Hatfield Associates Inc. Model 
5.0a was employed to estimate the forward looking economic costs”  Ex. 251T at 
2 (Spinks Direct).  Staff believes that the Hatfield model, as run with the 
inputs recommended by Staff, provides the most accurate estimate of the cost of 
providing service in high-cost areas.
 Staff, however, has not recommended that the Commission “select” a cost model 
in this proceeding. As Mr. Spinks noted, Staff does not believe that selection 
of a cost model is a central issue in this case.  Tr. at 1168-1169.  The 
ultimate issue for the Commission to determine, rather, is the cost of service 
in high-cost areas.  Furthermore, Mr. Spinks explained on cross examination 
that “the development of the models is an ongoing process.”  Tr.1171. 
 In particular, Staff does not believe it is either necessary or desirable to 
“select” one model over others, if the result of that selection would be to 
preclude the Commission from relying on other models in future proceedings (if, 
for example, other models are improved and shown to be superior), or if that 
selection were to give one model an automatic presumption of preferability in 
such proceedings.  Staff believes that the Commission must remain open to all 
models, as they continue to be revised and improved.
 Finally, as a policy matter, it is not necessary for the Commission to choose 
between the cost models in order to determine how much it costs to provide 
service in high cost areas of the state. This is so because the most critical 
factor in determining estimates is the choice of model inputs, rather than the 
models themselves.  Mr. Spinks explained:
[E]ven though USF support estimates vary from $30 to over $140 million, the 
Commission can considerably narrow that range by requiring adherence to its 
decisions regarding cost estimation methodology.  In other words, much of the 
variation in the cost estimates is more likely due to the use of inputs and 
factors the Commission has rejected in the past, than any other cause.

Ex. 260T at 8 (Spinks Reply).

D. Benchmark/Number of lines to support/Geographic level for calculation of 
costs

 1. Benchmark
 An explicit universal service mechanism would provide sufficient support in 
each high-cost area to allow telephone companies there to provide service at 
rates that are both affordable and comparable to rates charged in other areas.  
The level of support should be equal to the difference between the cost of 
providing service and the rates that are affordable and comparable.
 While it is useful for the Commission to understand the context for its 
decision in this adjudication, it need make no decision here about the level or 
nature of the affordable and comparable benchmark.
  2. Number of lines to support
 The Commission’s determination in this adjudication will provide information 
that may help policy makers decide whether a universal service mechanism should 
provide support to all lines in high-cost areas or to some subset of lines.  
Some have argued that only a single line should be supported and that only 
residential lines should be supported.  The Commission needs to determine the 
cost of supporting all lines versus primary residential lines, but it need make 
no decision in this proceeding about what lines should be supported.

 3. Geographic level for calculation of costs
 The Commission should determine costs at the smallest level of geography for 
which it has reliable information.  That is not to say that the level of 
support should be calculated at that level, because it may be impractical or 
unnecessary to vary the level of support among customers in an area.
 As discussed earlier, the development of the cost models is an ongoing 
process. Testimony and cross examination of witnesses in this proceeding have 
shown that neither model accurately locates customers in rural areas in all 
cases.  The lack of accurate locations results in uncertainty in the 
determination of plant requirements and subsequent costs. That result is 
unfortunate because it is precisely the rural areas that are associated with 
the high costs which require universal service support. 
 Because of the inability to precisely identify customer locations, the smaller 
the geographic area being estimated, the larger the potential for error. But as 
the area over which costs are being estimated increases, the amount of 
estimation error declines. Hence, Staff believes the minimum geographic level 
at which costs should be calculated is the wire center.
 Staff opposes the proposal of AT&T witness Ms. Baker to analyze costs over 
each company’s study area, see Ex. 371T at 16 (Baker Responsive), to the extent 
the proposal means that USF support would be made available on the same basis. 
If that were the case, USF support would become attached to customer locations 
in non high-cost locations, with insufficient support associated with the 
actual high cost loops. As stated in the report to the Commission on the FCC 
petition to disaggregate study area support: We want companies to compete for 
customers, not USF support. 
E. Surcharge
 The Commission need make no determination in this proceeding about whether 
disclosure of the benefits and costs of any explicit universal service 
mechanism should be required, prohibited, or otherwise regulated.  That issue 
is being addressed in the rulemaking and would likely be addressed in any 
legislation establishing a state mechanism, and the outcome has no effect on 
the cost of providing telephone service in high-cost areas.
F. Impact of FCC activities
 The only federal activity that might have a bearing on this proceeding is the 
FCC’s work on estimating the cost of providing telephone service in high-cost 
areas.  The FCC and the Universal Service Joint Board have undertaken that 
effort in parallel with more focused (but less platform-oriented) work at the 
states.  The current versions of both the BCPM and HAI models incorporate 
improvements driven by the work at the federal level.  This record includes 
information on the more important issues that have been debated there.  Thus 
the federal activity contributes to the Commission’s effort in this 
proceeding.  Beyond considering the record evidence in making its 
determination, the Commission need take no further account of federal 
activities.
G. Size of the fund/Reliability of fund size estimate
 This proceeding is about the cost of providing service, not the size of a 
universal service fund.  To the extent that parties make different assumptions 
about the level and nature of a benchmark for affordable and comparable service 
or the types of service that will be eligible for support, calculations of fund 
size will differ even where the cost of providing service is similar.
H. Should a “wireless cap” be implemented? If so, what is the appropriate 
investment cap?

 The use of the wireless investment cap is a conceptually sound model option 
designed to enable the proxy models to be consistent with the TELRIC 
methodology.  The Hatfield Model 5.0a documentation describes the addition of 
the “Investment Cap to Reflect Potential Wireless Technologies” as being 
“requested in the FCC’s FNRPM” and permitting “the specification of a 
user-adjustable cap on the model’s relevant wireline investments to reflect 
potentially more economical wireless distribution technologies.” (HAI Model 
Release 5.0a Model Description at 44.)
 The Staff cost estimates in this case used the wireless cap option in order to 
reflect the cost minimization principles underlying TELRIC.  Staff did not 
engage in an in-depth analysis of the wireless cap model mechanics or costs in 
this case, but rather, used the option because it is conceptually sound to 
invoke such an option if it exists in order to develop cost estimates 
consistent with TELRIC principles.

 I. If a revenue benchmark is adopted, what revenues should be included?

 This issue is being considered in the rulemaking proceeding.  Assuming that 
the Commission adopts a benchmark for affordable and comparable service that 
includes revenues from services such as access and vertical features that make 
use of the basic local network, then the cost of those services should also be 
included in the determination of costs.  In practical terms, there is very 
little additional cost.

 J. Should other sources of support be considered when estimating the size of 
the fund?

 This issue is being considered in the rulemaking proceeding.  The Commission 
need make no decision here on this issue.
K. Should households or housing units be used to estimate the cost of service?

 The cost of service should be determined based on actual lines in service 
within a given geographic area as opposed to using households or housing 
units.  Household data which is used in the Hatfield model is modified by 
actual line count data in order to build a network to serve all existing 
customers. The wirecenter template used by staff to develop cost estimates 
resizes the line counts in each cluster within the wirecenter to reconcile the 
number of lines estimated based on households with the actual number of lines 
in the wirecenter.  Actual line counts provide the most accurate estimate of 
cost of service for each wirecenter.
IV. COST MODELS
A1. Structure mix/Structure sharing
a. Staff structure mix/US West structure mix
 The Staff cost estimates in this case used the HM 5.0a default values for 
structure mix consistent with the Eighth Supplemental order in Docket 
UT-960369, the generic cost case.  Staff recommends the Commission continue 
using the default values in this case.
 US West witness Ms. Lent sponsors a sensitivity analysis in which US West 
actual plant mix is substituted for the default values, stating:
US West’s actual percent of aerial facilities in Washington is the most 
reliable guide available to the aerial plant that would exist in a 
forward-looking network.   

Ex. 291T at 16 (Lent Responsive).  The company’s position appears to be that 
past plant construction practices are assumed to be consistent with the 
forward-looking least cost assumptions of the TELRIC methodology, regardless of 
the lack of any evidence to support the proposition.
 Even if the assumption were true, at a minimum, the company specified plant 
mix cannot be correct.  Ex 293, attached to Ms. Lent’s responsive testimony, at 
“Change 2: Plant Mix” shows a change in the aerial distribution plant mix for 
the 10,000+ density zone from 85% to zero, stating that the 85% “is clearly an 
unreasonable view of what a reconstructed network would look like in 
Washington.” Ex. 291T at 16 (Lent Responsive). 
 However, the aerial cable used in the Hatfield model 10,000+ density zone is 
intended to represent the extensive network of riser cable in high-rise 
buildings which are found predominantly in the densest urban zone.  It is 
literally impossible to serve customers on the 60th floor of a high-rise with 
solely underground or buried facilities.  Hence, Staff recommends the 
Commission not adopt the U S WEST proposal for plant mix. Staff also notes that 
US West’s tariff permits the company to establish the demarcation point for 
distribution facilities to each floor of high-rise buildings.
b. US West, GTE and Sprint structure sharing

 GTE did not use the structure sharing formula required by the Commission in 
Docket UT-960369.  Instead, GTE used percentages for poles, buried, and 
underground that Mr. Collins stated were company-specific data obtained from 
the company’s “joint use administrators.”  Tr. at 656-58.  In the Ninth 
Supplemental Order in Docket UT-960369, Appendix A, the Commission set forth 
the GTE-specific structure sharing table for use with the Hatfield and BCPM 
models.  Those values should also be used in this proceeding.
 Staff notes that the Ninth Supplemental Order did not set forth specific 
structure-sharing values for Sprint.  In this proceeding, Sprint calculated the 
cost of pole attachments based on its own experience.  The percentage assigned 
to telephone is lower than that used in the BCPM runs by the Commission in 
Docket UT-960369.  For buried and underground, Sprint used the model default 
values which also are not consistent with the ones used by the Commission in 
that docket.
 The Commission should require US West, GTE and Sprint to each use the 
structure sharing formula set forth in Appendix A to the Ninth Supplemental 
Order in Docket UT-960369.
A2. Cost of money and depreciation rates

Staff addresses these issues at part VI(4) of this brief.

A3. Should the proxy model’s estimates of U S WEST drop lengths be adjusted to 
reflect the company’s special study? If so, how?

 The proxy model estimates of U S WEST drop length should not be adjusted to 
reflect the company’s special study.  
 In the Eighth Supplemental Order in Docket UT-960369, the Commission strongly 
suggested that parties substitute results of studies for judgment in 
determining drop lengths. In response to Staff Data Request No. 3 in this case, 
U S WEST provided a special study of drop lengths.
 Staff reviewed the study and concluded that the study results could not be 
used in the proxy models because while the loop lengths were calculated by lot 
types “urban”, “suburban”, and “rural,” the cost estimates were calculated by 
wirecenter, and there was no information provided regarding the relationship 
between the lot types and the wirecenters.  Rather, one was left to assume that 
the sample observations used to develop average drop lengths by lot types would 
be applied to the corresponding density zone, even though there was no evidence 
that the observations used