VIII. Geographic Deaveraging

When setting UNE rates, the Commission has generally adopted loop rates by geographic zones rather than one statewide average rate. The process of establishing zone rates is termed "geographic deaveraging."

JC propose three zones based on the loop costs per wire center from HM 5.3. Verizon also proposes three zones based on costs per wire center, but it groups the wire centers in a different manner, resulting in different zone rates than those proposed by JC. Both methodologies attempt to minimize the differences between average zone costs and individual wire center costs.

JC contend their deaveraging methodology results in a reasonably uniform distribution of wire centers among the three zones, whereas Verizon's methodology is designed to load the majority of lines in a single zone while leaving a handful of wire centers with the highest costs isolated in very high cost zones. (JC, 11/9/04, p. 83.)

In contrast, Verizon claims HM 5.3 zone estimates are based on a flawed deaveraging methodology that ignores higher cost wire centers and results in relatively few wire centers in the lowest cost zones. (Verizon, 8/6/04, p. 90.) Verizon claims JC's deaveraging proposal increases the likelihood of economically inefficient rates because the approach is biased towards minimizing the deviations in Zone 1 at the expense of greatly increased deviations in the other zones. (Id., and Verizon/Tucek, 8/6/04, p. 9.)

We reviewed the deaveraging methodologies of both JC and Verizon. The JC's method appears to group wire centers into the three zones based on an arbitrary cut-off of loop costs per wire center. Moreover, JC provide little, if any, explanation of their wire center groupings. While JC's methodology results in a fairly uniform distribution of wire centers between the three zones, the groupings appear to ignore the natural breaks between low and high cost wire centers that appear when the wire center cost results are graphed. In effect, the uniform groupings suggested by JC bias the Zone 3 rate downward.

Verizon provides a lengthier description of how it grouped wire centers into three zones based on a statistical technique to minimize the deviations in the zone. (Verizon/Tucek, 8/6/04, pps. 7-9.) The zones that result from Verizon's methodology group a few of the highest cost wire centers into Zone 3, with the bulk of wire centers grouped into Zones 1 and 2. As a result, Verizon's methodology leads to a significantly higher Zone 3 rate than JC's method.

We find the method proposed by Verizon better minimizes the deviations in wire center costs. It is reasonable to group only a few of the highest cost wire centers into Zone 3, rather than create more uniformly grouped zones that show a wider dispersion in the wire center costs in that zone and artificially lowers the zone's average rate. While we are not using the Verizon model to set rates, we can use Verizon's methodology to create zones based on the cost results from our HM 5.3 model run. We graphed the wire center costs for our run of HM 5.3, and found that two wire centers had significantly higher costs and were outliers on the graph of average wire center costs. (See Appendix C of this order which includes a graph of zone rates and a list of Verizon's wire centers by zone.) We conclude it is reasonable to isolate these two outlier wire centers into a fourth zone, rather than the three suggested by Verizon. This avoids these two high cost wire centers skewing the Zone 3 rate. Thus, our resulting zones are similar, but not identical to, the groupings suggested by Verizon. The table below shows UNE rates by zone for basic and DS-1 loops.

Table 7

Loop Rates by Deaveraged Rate Zones

	Zone 1	Zone 2	Zone 3	Zone 4
Basic Loops	$12.07	46.96	135.04	524.22
DS-1 Loops	76.49	170.02	297.77	678.08

We will not adopt deaveraged rates for DS-3 loops because we are not confident that either party's deaveraging methodology works appropriately for DS-3 loops. When we tested both Verizon's and JC's deaveraging methodologies on DS-3 loops, our results indicated a significantly higher cost in Zone 1 than in Zone 3. These results were illogical. Due to these deaveraging difficulties, we will adopt one average rate for DS-3 loops, as Verizon itself proposed.

Finally, it should be noted that Verizon's interim UNE rates were divided into two geographic zones, rather than the four zones we adopt today. As a result, interim loop rates by zone do not match up directly with permanent loop rates by zone. This will make it difficult to true-up interim to final UNE rates. Therefore, in order to implement the billing adjustment, discussed further in Section IX below, Verizon should use the percentage change in statewide average rates rather than zone rates when calculating any billing adjustments.