An Assessment of the Operational and Financial Health of Rate-of-Return Telecommunications Companies in more than 700 Study Areas:

An Assessment of the Operational and Financial Health of Rate-of-Return Telecommunications Companies in more than 700 Study Areas: 2007-2012 Harold Furchtgott-Roth Kathleen Wallman December 2014

Executive Summary Last year, we released a paper 1 on universal service and inter-carrier compensation and the possible effects of the FCC s 2011 Report and Order on universal service and intercarrier compensation. 2 Among many findings, the paper concluded that empirical information can help assess the effectiveness of government programs that aim to assist telecommunications companies to provide services in rural areas, consistent with Section 254 of the Communications Act. In this paper, we examine the operational and financial health of rural rate-of-return companies based on a six-year NECA database. Our data set comprises 705 study areas, and we look at available measures of company health and operations. The years covered in the data set allow us to examine data both before and after the issuance of the FCC s 2011 Report and Order. 3 Our observations collectively indicate that the vitality of rural rate-of-return companies has declined in the period 2011-2012 as compared to the period 2007-2010. We do not have a way of attributing this decline to the implementation of the FCC s 2011 Report and Order. We observe that the decline is consistent with the hypothesis that the Report and Order had an adverse effect on the health and viability of the companies. The measurements we studied were as follows: Loops; Gross telecommunications plant in service, 4 a measure of the gross value of plant and equipment that have been deployed, without consideration of depreciation; Gross telecommunications plant in service per loop; Total telecommunications plant in service minus accumulated depreciation, 5 a measure of the net value of plant and equipment; Total telecommunications plant in service minus accumulated depreciation per loop; Net plant, 6 a measure of the financial health of the company; 7 and 1 H. Furchtgott-Roth and K. Wallman, A Framework to Assess the FCC s 2011 Report and Order on Universal Service and Intercarrier Compensation, November 2013. 2 FCC, Report and Order and Further Notice of Proposed Rulemaking, ( 2011 Report and Order ), released November 18, 2011, FCC Rcd 177663 18414. 3 FCC, Report and Order and Further Notice of Proposed Rulemaking, ( 2011 Report and Order ), released November 18, 2011, FCC Rcd 177663 18414. 4 Field DL-160. 5 Field DL-190. 6 Field DL-220. 7 Net plant is equal to telecommunications plant and equipment in service plus cash and cash equivalents minus depreciation and amortization and other liabilities of the company. To the extent it captures all of the assets and liabilities of the company, this balance sheet concept is roughly equivalent to shareholder equity. 2

Net financial assets The Commission s stated objective in its 2011 Report and Order was to boost broadband adoption without adverse effects on the existing rural telecommunications service. As detailed below, the observations in each of these areas demonstrate the continuation of adverse changes throughout the period 2007-2012 and in some but not all cases, changes that are more markedly adverse in the period 2011-2012 versus the period 2007-2010. For example: Loop Declines: The mean number of loops in the 705 study areas declined each year in the 2007-2012 period, from roughly 5600 in 2007 to roughly 4400 in 2012, a decline of approximately 20%. While the average rate of loop decline attenuated over the period, 6.2% between 2007 and 2010 to 3.8% between 2010 and 2012, smaller study areas appear to have been hardest hit by the declines. Gross Plant in Service Per Loop Changes: As the number of loops declines, the gross plant in service spread across that smaller base increased from 2007 to 2012 as would be expected. The data show that telecommunications companies have slowed the rate at which they invest in supporting those loops. The average growth rate of gross plant in service per loop was at least 7.44% from 2007-2010. For the period, the average growth rate was 7.85%. After 2010, the growth rate declined to 6.02% in 2011 and 4.13% in 2012 with an average for the period of 4.85%. Declines in Net Financial Assets: The mean value of changes in net financial assets increased from 2007-2009, declined to $34,000 in 2010, and then declined further in 2011 and 2012 reaching a negative $206,000 in 2012. The median study area values declined throughout the period 2007-2012. This decline is a negative indicator of telecommunications companies wherewithal to invest in their networks. Even those measures that appear facially neutral or slightly positive mask negative changes affecting the smallest study areas. In sum, we found no optimistic news in the 2012 NECA data regarding the health and vitality of telecommunications companies serving Americans living in rural areas. Introduction Last year, we released a paper 8 on universal service and inter-carrier compensation and the possible effects of the FCC s 2011 Report and Order on universal service and intercarrier compensation. 9 Among many findings, the paper concluded that empirical 8 H. Furchtgott-Roth and K. Wallman, A Framework to Assess the FCC s 2011 Report and Order on Universal Service and Intercarrier Compensation, November 2013. 9 FCC, Report and Order and Further Notice of Proposed Rulemaking, ( 2011 Report and Order ), released November 18, 2011, FCC Rcd 177663 18414. 3

information can help assess the effectiveness of government programs that aim to assist telecommunications companies to provide services in rural areas, consistent with Section 254 of the Communications Act. In March 2014, the FCC s Wireline Competition Bureau issued a report that presents financial information from a sample of rate-of-return telephone companies. 10 The report provides an optimistic assessment of the financial condition based on its assessment of a 50 company sample. 11 The report is based on information that is not publicly available, 12 and we consequently can neither replicate its results nor or comment on the reasonableness of the Bureau s assessment. In this paper, we examine in more detail the operational and financial health of rural telecommunications companies in rural areas both before and after the 2011 FCC Order. We employ a database of six years of operational information assembled by NECA from 705 study areas for rate-of-return carriers to assess the operational and financial health of the telecommunications companies serving those areas. 13 The most recent NECA information is for calendar year 2012. In 2012, these study areas served more than 3 million loops, or, assuming approximately one loop per household, approximately 3% of U.S. households. We focus on the following seven combinations of information from the NECA database: 14 10 FCC Wireline Competition Bureau, Docket 10-90, Universal Service Implementation Progress Report, March 18, 2014. 11 Ibid., p. 15. 12 Form 481 financial data. 13 We have prepared and analyzed a database of information for rate-of-return carriers. The database consists of information reported ly between 2008 and 2013 by NECA to the FCC. The information can be found under Universal Service Fund Data at the FCC website http://transition.fcc.gov/wcb/iatd/neca.html. Each submission consists of information for the prior calendar year; thus, the database reflects information from 2007-2012. Many of the column headings can be found at links from http://galbithink.org/telcos/nts-cost-data.htm. We have translated all dollar values into 2014 1 st quarter dollars using the GDP deflator. We included only those study areas that have records in each year. We also excluded two study areas that are not in rural areas, both owned by Cincinnati Bell. We also excluded four study areas that had negative values for gross plant minus depreciation, or net plant. These four study areas are Perkinsville Telephone Company in Vermont, Calhoun Telephone Company in Mississippi, Lakeside Telephone Company in Mississippi, and Asotin Telephone Company in Oregon. Negative values in these study areas lead to difficult interpretations of growth. Had we included these companies, the results would paint an less healthy picture of rural telephone companies than the one we find. The net result is a database of 705 study areas. 14 Each of the records in the NECA database has more than 120 fields of information. 4

Loops; Gross telecommunications plant in service, 15 a measure of the gross value of plant and equipment that have been deployed, without consideration of depreciation; Gross telecommunications plant in service per loop; Total telecommunications plant in service minus accumulated depreciation, 16 a measure of the net value of plant and equipment; Total telecommunications plant in service minus accumulated depreciation per loop; Net plant, 17 a measure of the financial health of the company; 18 and Net financial assets For each type of information, we find that rural telecommunications companies in the majority of study areas had operational and financial conditions that were weakened in 2011 and 2012. 19 These results are consistent with there being an adverse effect from the FCC s 2011 Report and Order, although we have no formal test for causation. For none of information reviewed in this paper do we find appreciable improvement in operational and financial conditions of the rural telecommunications companies after the FCC s 2011 Report and Order. This paper finds that rural telecommunications communications companies are weaker after 2010, a finding that will surprise no one who follows the industry. We assign no causation to the declining conditions. Nor do we examine or conclude that customers in these study areas have lesser service than customers in those areas where competing wireless and other forms of telecommunications may be available. But the deteriorating operational and financial health of wireline rural telecommunications companies serving rural areas is clear. We review the results for each information area below. 15 Field DL-160. 16 Field DL-190. 17 Field DL-220. 18 Net plant is equal to telecommunications plant and equipment in service plus cash and cash equivalents minus depreciation and amortization and other liabilities of the company. To the extent it captures all of the assets and liabilities of the company, this balance sheet concept is roughly equivalent to shareholder equity. 19 The FCC adopted the 2011 Report and Order on October 27, 2011 and released it on November 18, 2011. The potential effects of the 2011 Report and Order were likely known for much of the 4 th quarter of 2011, if not before. Consequently, if the 2011 Report and Order were to have a harmful effect on the financial characteristics of rural telephone companies, those characteristics in 2011 might fall below earlier levels. 5

Loops Loops are a standard measure of the commercial activity and a rough approximation of the customer base of a telephone company. 20 Generally, loop growth is associated with expansion of the customer base, and loop decline is associated with contraction of the customer base. Loop decline does not necessarily mean that there is less physical plant, merely that there are fewer customers. In Table 1, we present information on loops for the 705 study areas. 21 The first line of Table 1 displays the mean number of loops for the study areas in each year from 2007-2012. The mean number of loops declines each year, from roughly 5600 in 2007 to roughly 4400 in 2012, a decline of approximately 20%. The next five lines present in each year the maximum number of loops for a study area, the 90 th percentile of loops, the median or 50 th percentile of loops, the 10 th percentile of loops, and the minimum number of loops in a study area, respectively. We will use this format in each of the tables. With the exception of the minimum number of loops in a study area, all other statistics are consistently declining throughout the period. During the period, roughly 80% of study areas have between 500 and 13,000 loops. In each year, the mean number of loops is greater than the median number of loops, indicating that a small number of study areas have loops that are substantially greater than the median. The total number of loops for the 705 study areas was nearly 4 million in 2007 but consistently declined to slightly more than 3.1 million by 2012. Still, throughout this period, the 705 study areas contained at least 2.1% of the switched access lines in the United States in 2007 and steadily increasing to 3.2% of switched access lines in 2012. 22 The number of switched access lines in the 705 study areas declined less rapidly than in the remainder of the United States. 20 A loop can support multiple services. 21 Field 060 in the NECA database. 22 Based on authors calculations. See FCC, Wireline Competition Bureau, Figure 1 in Local Telephone Competition: Status as of December 31, 2012 November 2013, at http://transition.fcc.gov/daily_releases/daily_business/2013/db1126/doc- 324413A1.pdf; FCC, Wireline Competition Bureau, Figure 1 in Local Telephone Competition: Status as of December 31, 2010 October 2011, at https://apps.fcc.gov/edocs_public/attachmatch/doc-310264a1.pdf; and FCC, Wireline Competition Bureau, Table 1 in Local Telephone Competition: Status as of December 31, 2007 September 2008, at https://apps.fcc.gov/edocs_public/attachmatch/doc- 285509A1.pdf. 6

Table 1 Loops for the 705 Study Areas 2007 2008 2009 2010 2011 2012 mean 5,641 5,354 5,072 4,830 4,610 4,418 Maximum 90,478 83,258 76,454 71,415 66,828 63,310 90th percentile 12,795 12,061 11,242 10,749 10,329 10,131 50th percentile 3,139 2,984 2,843 2,677 2,585 2,513 10th percentile 656 624 609 582 557 537 Minimum 16 18 21 20 32 33 total 3,976,553 3,774,514 3,575,842 3,405,405 3,250,321 3,114,506 A striking feature of Table 1 is the consistent decline in the number of loops. Table 2 shows the decline in the number of loops by comparing the number of loops in adjacent years. The decline in the mean number of loops declines from 285 from 2007-2008 to 192 from 2011 to 2012, but these declines in the number of loops are on a dwindling base. Although some study areas add loops, far more study areas are losing loops, sometimes thousands of loops in a year. More than 90% of study areas lost loops in each year. In total, the 705 study areas lost at least 135,000 loops in each year. Table 2 Decline in the Number of Loops for the 705 Study Areas 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean (287) (282) (242) (220) (193) Maximum 778 4,001 340 415 2,248 90th percentile (11) (14) (9) (8) (4) 50th percentile (125) (130) (105) (98) (86) 10th percentile (628) (637) (585) (513) (501) Minimum (7,220) (6,804) (5,039) (4,587) (3,518) total (202,039) (198,672) (170,437) (155,084) (135,815) Table 3 shows similar information based on the ratio of the number of loops from year to year. The ratio of the mean number of loops from Table 2, the last line of Table 3, reveals a consistent decline of 5%, as the number of loops in each year for the entire 7

sample. The mean percentage of losses tends to smaller than the ratio of means, but the mean percentage loss is still at least 3.3% per year. Of course, in each year, some study areas show an increase in the number of loops, but more than 90% of study areas have loop loss in each year. Many study areas have substantial losses. More than 10% of study areas have loop losses exceeding 7.5% per year. Throughout this paper, we will focus on differences between the 2007-2010 period and the 2010-2012 period. The last two columns of Table 3 show the ized loop losses across these longer periods, which show less variation than the year-to-year changes. A study area that loses many lines one year might lose fewer lines the next year, and consequently the losses averaged over a few years are less pronounced than the year-byyear losses. In the last two columns, the average decline in the number of loops attenuates from 6.2% between 2007 and 2010 to 3.8% between 2010 and 2012. The ratio of the averages from Table 1, the last lines in the last two columns of Table 3, show less divergence indicating that much of the attenuation in line loss is for the smaller study areas. Although loop losses attenuate in the later period, it is difficult to conclude that changes in the number of loops has stabilized to a healthy situation after 2010 and the possible effects of the 2011 Report and Order. The situation for rural telecommunications companies as measured by line loss is poor in both periods. Table 3 Decline in the Ratio of the Number of Loops for the 705 Study Areas 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean -4.51% -3.27% -4.16% -3.32% -3.89% Maximum 65.52% 801.80% 47.18% 470.00% 35.63% 90th percentile -0.97% -1.43% -1.06% -0.82% -0.59% 50th percentile -4.43% -4.54% -4.08% -3.94% -3.88% 10th percentile -8.65% -8.39% -7.84% -7.64% -7.80% Minimum -41.74% -27.38% -32.02% -21.46% -32.65% ratio of means -5.08% -5.26% -4.77% -4.55% -4.18% 2007-2010: -6.22% 200.61% -2.88% -6.53% -11.36% -23.61% -5.04% 2010-2012: -3.79% 129.49% -1.13% -3.97% -7.39% -22.04% -4.37% For each measure of operational and financial health, we also look at the percentage of study areas achieving specific growth or loss levels. We focus on three: 2.5% growth; no growth; and 2.5% loss. The 2.5% growth target is roughly the real growth rate of the American economy over a long period of time. Study areas growing more rapidly than 2.5% ly are growing more rapidly than the American economy. These might be deemed economically healthy study areas. Study areas growing between 0% and 2.5% 8

ly are growing but more slowly than the American economy. These might be described as slow growing study areas. Study areas declining less than 2.5% per year might be characterized as slowly declining study areas. Study areas declining by more than 2.5% ly can be described as rapidly declining study areas, ones that are likely under substantial operational and financial stress. In Table 4, we present the percentile rank for loop growth of the 705 study areas. Fewer than 4% of study areas meet the 2.5% growth target in any year. For the longer periods, fewer than 2% meet the 2.5% growth target. Even the 0% growth target does little better. Fewer than 3% meet the target in the 2007-2010 period, and fewer than 5% meet the target between 2010 and 2012. More than 91% of study areas fail to even meet the 2.5% decline level between 2007 and 2010, and 75% fail to meet that target from 2010-2012. In terms of loop loss, the vast majority of study areas are in operational and financial distress. Table 4 Percentile Rank for the Rate of Growth of the Number of Loops for the 705 Study Areas Rate of Growth 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2.50% 0.973 0.98 0.983 0.983 0.96 0.00% 0.934 0.963 0.94 0.933 0.917-2.50% 0.772 0.816 0.763 0.756 0.708 2007-2010: 0.982 0.974 0.918 2010-2012: 0.985 0.954 0.756 The number of loops is falling in all periods. The losses are slightly attenuated in the later period, but losses are still substantial. It would be impossible to view any of the tables on loop loss and infer that the industry is healthy. Of course, some consumers may be substituting wireless and other telecommunications services for the wireline services of the telephone company, but it is impossible to review the information on loops and find any interpretation that reflects telephone company health and growth. Gross Telecommunications Plant In Service Gross telecommunications plant in service is a measure of the gross value of plant and equipment for each of the 705 study areas. It is a measure of gross investment and does not take into account depreciation, which is reviewed in the next section. It includes both central office equipment as well as the distribution network of fiber and copper loops. Until they are taken out of service, all plant and equipment are presumably recorded in this field. The change in gross plant in service from one year to the next should reflect investment in new plant and equipment less plant and equipment taken out of service. 9

One would expect for a healthy and growing business this change in gross plant and equipment to be positive from one year to the next. 23 In Table 5, we present summary statistical information on the value in 2014 Q1 dollars of total gross telecommunications plant in service for all 705 study areas for the years 2007 2012. The first row of Table 5 shows that the mean value of gross telecommunications plant in service ranged between $32 million and $36 million per study area, increasing slightly throughout the period. The maximum value for a study area was nearly $400 million, and the smallest value was less than half a million dollars. The median value of gross telecommunications plant in service shows the same pattern and ranged between $19 million and $21 million. Approximately 80% of study areas had gross plant in service of between $4 million and $43 million. Total gross plant in service for all study areas ranged between $22 billion and $26 billion. Table 5 Gross Telecommunications Plant In Service: 2007-2012 (in 2014 Q1 dollars) 2007 2008 2009 2010 2011 2012 mean 32,157,040 33,082,293 34,346,077 35,348,187 35,812,026 35,855,641 Maximum 373,968,424 386,577,272 385,656,334 383,471,947 388,007,551 391,878,421 90th percentile 37,580,678 39,490,402 41,193,766 41,728,701 42,836,605 42,556,611 50th percentile 19,191,199 19,761,754 19,738,210 20,252,386 20,745,905 20,870,596 10th percentile 4,558,368 4,745,220 4,865,629 4,811,655 5,023,268 4,936,606 Minimum 384,106 421,078 453,527 448,083 439,451 438,949 total 22,670,713,255 23,323,016,477 24,213,984,408 24,920,471,870 25,247,478,271 25,278,227,095 Source: NECA databases and authors' calculations On the surface, these results might suggest a healthy investment environment in the study areas with increasing plant in service each year. But the aggregate increases mask important differences in the pattern of changes in gross plant over time. In Table 6, we present in 2014 Q1 dollars the difference in gross plant and equipment between adjacent years. These changes in gross plant in service are roughly a measure of how much additional plant a telecommunications company puts in service each year, assuming that none is taken out of service. If a telephone company were healthy and expanding, one would expect robust growth in gross plant in service each year. 23 One can construct hypothetical counterexamples. Thus, a $10,000 router might hypothetically replace a less capable $100,000 switch. But for other types of plant and equipment, one would not expect to see this pattern. 10

As can be seen in the first row of Table 6, the average growth in gross plant in service was between $900,000 and $1.3 million ly between 2007 and 2010. The average increase in gross plant in service declines precipitously to $464,000 in 2011 and barely $43,000 in 2012. The 90 th, 50 th, and 10 th percentiles all show a similar pattern of dramatic decline in gross plant in service after 2010. The median increases in gross plant are between $156,000 and $197,000 from 2007-2010. By 2011, the median increase falls to just $23,000. By 2012, the median increase is a decline of $45,000. The declines in gross plant in service are quite large for the bottom 10% of study areas. Between 2007 and 2010, fewer than 10% of study areas lost more than $736,000 in gross plant in service in any year. For 2011, 10% of study areas lost more than $1.059 million. By 2012, more than 10% of study areas lost more than $1.837 million. Table 6 Change in Gross Telecommunications Plant In Service: 2007-2012 (in 2014 Q1 dollars) 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean 925,253 1,263,784 1,002,110 463,839 43,615 Maximum 36,380,335 64,251,018 38,196,019 15,663,073 24,224,743 90th percentile 1,237,034 1,205,944 1,248,842 753,697 381,050 50th percentile 155,920 197,104 171,273 22,785 (45,301) 10th percentile (709,854) (586,216) (735,517) (1,059,532) (1,837,120) Minimum (14,795,355) (7,569,631) (17,745,474) (37,106,579) (22,025,794) Source: NECA databases and authors' calculations In Table 7, we present the percentage change in gross plant in service for the 705 study areas. The mean across the 705 study areas of the change in plant in service, the first row of Table 7, shows a pattern that remains relatively high through 2011 and then declines precipitously in the last year from 2011-2012. The mean values remain high, above 2.85%, through 2010. In 2011, the mean value declines to 1.75%, and then turns negative in 2012. For the period 2007-2010, the mean increase in gross plant and equipment is 3.19%. That rate of increase declines to 0.67% from 2010-2012. The median value in the change in gross plant in service, the fourth line of Table 7, shows a similar pattern. For the median value, half of the 705 study areas have a faster growth rate in gross plant in service, and half a slower growth rate. The median growth rate is at least 0.92% through 2010. Then the median growth rate slows to 0.17% in 2010-2011, and turns to a negative 0.44% in 2011-2012. The median value for growth of gross 11

plant and equipment in the period 2007-2010 is 1.44%, which declines to a contraction of 0.04% between 2010 and 2012. The final row of Table 7 shows the change in the mean value of gross plant in service from the mean values in Table 5. This is a measure of the change in the total plant in service in the 705 study areas. The change in the mean value is at least 2.88% ly between 2007 and 2010 and then declines to 1.31% and 0.12% after 2010. For the entire 2007-2010 period, the change in mean value is 3.20% which declines to 0.06% from 2010 to 2012. The change in gross plant between 2011 and 2012, whether measured as the mean across 705 study areas, or ratio of the means from Table 5, is close to zero. This change is dramatically less than that reported by the FCC s Wireline Competition Bureau for changes in Telephone Plant in Service of 6.6%. 24 Table 7 Percentage Change in Gross Telecommunications Plant In Service: 2007-2012 (in 2014 Q1 dollars) 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2007-2010: 2010-2012: mean 3.22% 5.13% 2.85% 1.75% -0.06% Maximum 104.47% 848.78% 68.72% 82.72% 126.28% 90th percentile 4.88% 4.97% 4.98% 3.05% 2.03% 50th percentile 1.07% 1.25% 0.92% 0.17% -0.44% 10th percentile -3.67% -2.95% -3.11% -3.34% -7.25% Minimum -34.03% -29.42% -31.39% -27.73% -38.24% change in mean value 2.88% 3.82% 2.92% 1.31% 0.12% Source: NECA databases and authors' calculations 3.19% 137.21% 4.49% 1.44% -1.72% -13.32% 3.20% 0.67% 47.91% 2.33% -0.04% -4.05% -19.86% 0.06% In Table 8, we present the percentile rank of study areas that have a certain growth rate in gross plant and equipment. To attain a growth rate of 2.5%, the same as the American economy, places a study area in the top 38% of study areas from 2007-2010. But such a growth rate would place the study area in the 23% of study areas from 2010-2012. Only 30% of study areas failed to grow in gross plant and equipment from 2007-2010; fully half of study areas failed to grow from 2010-2012. Only 6% of study areas declined at 24 FCC Wireline Competition Bureau, Docket 10-90, Universal Service Implementation Progress Report, March 18, 2014, Figure 8. It is not clear whether the Bureau was referring to gross or net Telephone Plant in Service. 12

more than 2.5% for gross plant and equipment from 2007-2010; more than 17% declined at a greater rate from 2010-2012. Table 8 Percentile Rank for the Rate of Growth of Gross Plant and Equipment for the 709 Study Areas (in 2014 Q1 dollars) 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2.50% 0.599 0.603 0.638 0.715 0.778 0.00% 0.375 0.341 0.376 0.469 0.556-2.50% 0.138 0.117 0.121 0.138 0.229 Source: NECA databases and authors' calculations 2007-2010: 0.615 0.303 0.061 2010-2012: 0.763 0.506 0.174 For most study areas, gross plant and equipment are not growing at a healthy rate throughout the period 2007-2010. The value of gross plant in service grows more slowly after 2010, consistent with a negative effect of the 2011 FCC Report and Order. Gross plant in service per loop Generally, we find that gross plant and equipment per loop are increasing from 2007-2012. With the number of loops declining steadily and gross plant increasing slowly, it is not surprising that gross plant per loop is increasing over time. If one were examining a newly built, greenfield telecommunications company, gross plant per loop might be considered a measure of the quality of service provided by the company. But for an existing telecommunications network, gross plant per loop may have little direct relationship to the quality of the plant. For example, consider a company with a fixed amount of plant in service. The company makes no additional investments in plant, but the company loses half of its customer loops one year. The company will report having twice the plant in service per loop the following year even though the actual amount and quality of the plant in service has not changed; indeed, considering depreciation as we do in the next section, the quality of plant has actually declined. The combination of increases in gross plant and decreases in loops often means that fewer customers are left to support and pay for the same network. In Table 9, we present the telecommunications plant in service per loop for the 705 study areas. This information is taken from the underlying information for Tables 1 and 5. Table 9 reveals that plant in service per loop increases consistently throughout the period with mean values beginning at $7,740 in 2007 and ending with $10,127 in 2012. Although there are outliers reflected with the maximum and minimum values in Table 9, 80% of study areas had gross plant in service per loop of between approximately $3,400 and $17,100. 13

Table 9 Telecommunications Plant in Service per Loop for the 705 Study Areas (in 2014 Q1 dollars) 2007 2008 2009 2010 2011 2012 mean 7,740 8,235 8,861 9,471 9,831 10,127 Maximum 149,568 111,708 114,912 115,673 118,183 110,634 90th percentile 11,908 13,226 14,196 15,606 16,682 17,085 50th percentile 5,861 6,365 7,012 7,640 8,042 8,435 10th percentile 3,418 3,719 4,068 4,268 4,429 4,588 Minimum 1,727 2,072 2,057 2,147 2,058 1,965 Table 10 presents the changes by year in telecommunications plant in service per loop. The mean value of those changes per study area between 2007 and 2010 is between $495 and $626. After 2010, the mean value of the change in plant per study area drops to $360 in 2011 and $296 in 2012. Generally, most study areas have increasing gross plant per loop. Table 10 Changes in Telecommunications Plant in Service per Loop for the 705 Study Areas (in 2014 Q1 dollars) 2008-2007 2009-2008 2010-2009 2011-2010 2012-2011 mean 495 626 609 360 296 Maximum 21,370 18,566 14,139 7,598 13,829 90th percentile 1,239 1,401 1,532 1,418 1,128 50th percentile 352 404 377 339 296 10th percentile (36) 21 (18) (84) (490) Minimum (37,860) (16,671) (9,118) (57,836) (20,969) Table 11 shows similar information with the decline in the percentage growth rate of telecommunications plant per loop. The average growth rate was at least 7.44% 14

from 2007-2010. For the period, the average growth rate was 7.85%. After 2010, the growth rate delines to 6.02% in 2011 and 4.13% in 2012 with an average for the period of 4.85%. The same pattern of decline in the latter period holds for the ratio of means from Table 9. Table 11 Decline in the Growth Rate of Gross Plant per Loop 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean 8.27% 8.83% 7.44% 6.02% 4.13% Maximum 122.23% 155.95% 72.45% 94.00% 139.59% 90th percentile 18.82% 17.56% 16.19% 13.47% 11.27% 50th percentile 6.13% 6.72% 5.91% 4.86% 4.23% 10th percentile -0.41% 0.44% -0.25% -1.29% -4.77% Minimum -29.29% -22.39% -30.55% -75.63% -26.30% ratio of means 6.39% 7.60% 6.88% 3.80% 3.01% 2007-2010: 7.85% 54.67% 14.29% 6.86% 2.30% -11.61% 6.96% 2010-2012: 4.85% 62.42% 11.09% 4.62% -1.70% -49.09% 3.41% In table 12, we present the percentile rank of study areas that have a certain growth rate in gross plant and equipment per loop. Only 11% of study areas did not obtain a growth rate of 2.5% in gross plant and equipment per loop from 2007-2010. In contrast, 29% of study areas did not obtain such a growth rate from 2010-2012. Fewer than 5% of study areas failed to grow in gross plant and equipment per loop from 2007-2010; more than 14% of study areas failed to grow from 2010-2012. Only 2.2% of study areas declined at more than 2.5% for gross plant and equipment per loop from 2007-2010; more than 7% declined at a greater rate from 2010-2012. Table 12 Percentile Rank for the Rate of Growth of Gross Plant and Equipment for the 705 Study Areas 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2.50% 0.214 0.18 0.252 0.26 0.345 0.00% 0.103 0.09 0.104 0.132 0.221-2.50% 0.071 0.045 0.056 0.063 0.142 Source: NECA databases and authors' calculations 2007-2010: 0.112 0.045 0.022 2010-2012: 0.291 0.143 0.074 Generally, gross plant and equipment per study area grew from 2007-2012 but the rate of increase declined after 2010. The growth of this statistic reflects more the decline of loops rather than greater investments by telephone companies. 15

Net plant in service Thus far, we have examined gross plant in service. In this section, we review net plant in service, which is equal to gross plant in service minus depreciation. Net plant in service is a more accurate measure of the value of plant in service as all plant and equipment, particularly in a rapidly technologically changing industry such as telecommunications, depreciates. Table 13 presents the net plant in service of the 705 study areas. The net plant in service in Table 13 tends to be approximately only one third as large as the gross plant in service in Table 5. That is, for most study areas, most of the value of gross plant has depreciated. In all years, the mean net plant in service is between $12.0 million and $12.6 million. There is enormous range in the value of net plant in service. The largest study areas have approximately $200 million in net plant; the smallest study areas have less than $100,000 in the value of net plant. 25 Eighty percent of study areas have net plant values between $1 million and $15.3 million. The mean net plant in service increases from 2007 to 2010 and then declines from 2010 to 2012. Most statistics in Table 10 show a similar pattern except 10 th percentile. The 10 th percentile of net plant in service declines consistently from 2007 2012 indicating that net plant in the smallest study areas has been declining more rapidly and for a longer time than the net plant in larger study areas. net plant 2007 net plant 2008 net plant 2009 net plant 2010 net plant 2011 net plant 2012 mean 12,003,980 12,184,805 12,498,207 12,592,724 12,454,130 12,244,890 Maximum 162,038,088 171,641,674 184,307,534 201,068,846 192,015,172 197,816,722 90th percentile 14,616,838 14,753,015 14,736,863 15,042,243 15,149,108 15,272,679 50th percentile 6,496,269 6,577,889 6,474,459 6,687,846 6,539,449 6,167,486 10th percentile 1,469,072 1,407,014 1,320,520 1,247,990 1,167,296 1,045,141 Minimum 106,628 99,662 9,363 71,549 70,972 45,334 total 8,462,805,871 8,590,287,610 8,811,236,197 8,877,870,197 8,780,161,613 8,632,647,136 Source: NECA databases and authors' calculations Table 13 Net Telecommunications Plant In Service: 2007-2012 (in 2014 Q1 dollars) Table 14 shows the change in net plant in service in adjacent years. The mean value of net plant in service increases between 2007 and 2010 and then falls from 2010 to 2012. Curiously, the median and lower percentile values of the change in net plant are negative 25 We have excluded those study areas with negative values of net plant. 16

in all periods. That is, a majority of study areas are losing net plant throughout the period. In every period however, some study areas have growing net plant in service. Table 14 Change in Net Plant in Service for the 705 Study Areas 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean 180,825 313,402 94,516 (138,594) (209,240) Maximum 33,290,390 48,938,913 36,374,640 12,617,016 17,090,592 90th percentile 2,199,038 1,991,990 1,813,438 1,682,354 1,286,286 50th percentile (169,151) (204,824) (200,664) (272,725) (282,438) 10th percentile (1,249,194) (1,313,794) (1,404,884) (1,768,247) (1,957,535) Minimum (8,363,922) (13,735,495) (16,104,305) (14,926,339) (17,000,523) total 127,481,739 220,948,587 66,634,000 (97,708,584) (147,514,477) Table 15 displays the changes in the percentage growth rates of net plant in service. The mean of the growth rate is positive in all years except 2011-2012. The mean of this measure has substantial volatility for a few reasons, including: net plant in service can be small or even negative; and ratios in general can have volatility when the denominator is small. As one can see with the maximum value of the growth rate, the second line of Table 15, values can exceed 1000%. Still, the pattern is increases in the mean values of net plant for the first three years and then declines in the last two years. The average growth rate for the first three years is 1.11% and declines to -2.63% in the last two years. These values are consistently lower than the year-by-year averages. Notice that the last row of Table 15, the ratio of means from table 13, has the same pattern. A more stable and consistent measure of change of ratios is the median or the 50 th percentile. The median value of the percentage change in net plant in service is consistently negative throughout the years of observation and turns more negative in the last two years. The change in net plant between 2011 and 2012, whether measured as the mean across 705 study areas, or ratio of the means from Table 15, is either -0.81% or -1.61%. This change is dramatically less than that reported by the FCC s Wireline Bureau for changes in Telephone Plant in Service unclear whether gross or net of 6.6% growth. 26 26 FCC Wireline Competition Bureau, Docket 10-90, Universal Service Implementation Progress Report, March 18, 2014., Figure 8. 17

It is difficult to review any measures of net plant in service and reach any conclusion other than the 705 study areas have widespread declining net plant in service. This is not a healthy situation for the telephone companies operating in these study areas. Table 15 Percentage Growth of Net Plant in Service for the 705 Study Areas 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean 3.61% 11.49% 3.11% -1.09% -0.81% Maximum 610.07% 3960.90% 1415.53% 380.81% 1701.14% 90th percentile 21.14% 22.13% 21.11% 15.81% 12.96% 50th percentile -3.78% -5.23% -4.53% -6.24% -7.42% 10th percentile -15.78% -15.71% -15.21% -16.38% -18.72% Minimum -63.49% -98.25% -38.58% -53.28% -48.43% ratio of means 1.51% 2.57% 0.76% -1.10% -1.68% 2007-2010: 1.11% 173.50% 17.32% -2.95% -12.38% -54.07% 4.90% 2010-2012: -2.63% 259.53% 12.73% -6.18% -16.32% -49.50% -2.76% In table 16, we present the percentile rank of study areas that have a certain growth rate in net plant and equipment. The pattern shows a decline in the operational and financial stability of the study areas after 2010. Only 30% of study areas obtained a growth rate of 2.5% in net plant and equipment from 2007-2010. In contrast, even fewer, only 23% of study areas obtained such a growth rate from 2010-2012. More than 61% of study areas failed to grow in net plant and equipment per loop from 2007-2010; more than 70% of study areas failed to grow from 2010-2012. More than 51% of study areas declined at more than 2.5% for net plant and equipment per loop from 2007-2010; more than 62% declined at a greater rate from 2010-2012. Table 16 Percentile Rank for the Rate of Growth of Net Plant in Service for the 705 Study Areas Rate of Growth 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2.50% 0.694 0.715 0.709 0.766 0.768 0.00% 0.645 0.657 0.652 0.705 0.731-2.50% 0.552 0.58 0.585 0.645 0.675 2007-2010: 0.698 0.616 0.515 2010-2012: 0.765 0.703 0.629 Net plant in service per loop In this section, we examine changes in net plant in service per loop, much as we examined gross plant in service per loop earlier. In Table 17, we present information on 18

net plant in service per loop. The mean net value per loop ranges in value between $3300 and $3900, or approximately 40% of the values in Table 9 for gross plant in service per loop. There are substantial variations in these values from less than $100 to more than $129,000 per loop. 27 The 10 th percentile and the 90 th percentile provide a narrower range for the 80% of study areas in the middle, between $800 and $8,300 per loop. The mean value of net plant in service per loop increases from 2007-2010 and then declines slightly in 2011. The 90 th percentile and the median value increase consistently over time. In contrast, the 10 th percentile decreases or remains constant from 2008 onward. Table 17 Net Telecommunications Plant in Service per Loop for the 705 Study Areas (in 2014 Q1 dollars) 2007 2008 2009.0 2010.0 2011 2012 mean 3,389 3,499 3681.7 3841.6 3,815 3,838 Maximum 129,611 84,616 78328.7 86147.7 78,727 64,070 90th percentile 5,314 6,070 6548.1 7080.0 7,565 8,222 50th percentile 2,060 2,266 2356.5 2462.9 2,542 2,596 10th percentile 932 947 920.8 873.9 844 844 Minimum 268 199 3.6 57.2 43 31 Table 18 presents the changes in net plant in service per loop. The mean value increases at least $111 ly per loop from 2007-2010 but falls substantially afterwards, and even negative from 2010-2011. Although there is enormous variation between the study areas with the largest growth and decline in net plant in service, the median value is remarkably stable. The median increases slightly from 2007-2009 and declines afterwards. 27 We excluded study areas with negative net plant. 19

Table 18 Changes in Net Telecommunications Plant in Service per Loop for the 705 Study Areas (in 2014 Q1 dollars) 2008-2007 2009-2008 2010-2009 2011-2010 2012-2011 mean 111 183 160 (26) 23 Maximum 9,990 17,780 7,819 6,761 15,954 90th percentile 689 751 881 700 582 50th percentile 17 6 (9) (40) (58) 10th percentile (266) (306) (293) (393) (478) Minimum (44,994) (19,274) (12,074) (44,171) (18,928) Table 19 presents the percentage changes in the growth rate of net plant per loop. The results are similar to those in Table 15. Specifically, the maximum and minimum values are volatile changing by more than 1,000% per year. The average changes are generally increases, but the rate of increases diminishes substantially after 2010. For the period 2007-2010, the average increase was 5.6%. The average growth rate falls to 1.4% after 2010. The median increases slightly from 2007-2009 and declines afterwards. From 2009-2012, large numbers of study areas have declining net plant per loop. It is difficult to look at net plant per loop and not see that telephone companies in many study areas are declining in financial health. Table 19 Changes in the Growth Rate of Net Plant per Loop 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean 8.59% 15.09% 7.72% 3.05% 3.28% Maximum 615.53% 4202.25% 1484.58% 404.13% 1807.08% 90th percentile 28.25% 27.96% 26.80% 20.58% 17.18% 50th percentile 0.88% 0.22% -0.55% -1.94% -3.05% 10th percentile -12.19% -12.19% -11.83% -13.59% -15.26% Minimum -62.19% -98.19% -35.55% -80.72% -45.57% ratio of means 3.26% 5.22% 4.34% -0.68% 0.59% 2007-2010: 5.61% 190.68% 22.24% 2.21% -8.35% -51.97% 4.27% 2010-2012: 1.40% 294.80% 17.46% -1.86% -12.67% -57.58% -0.05% 20

In table 20 we present the percentile rank of study areas that have a certain growth rate in net plant and equipment per loop. The pattern shows a decline in the operational and financial stability of the study areas after 2010. Fully 48% of study areas obtained a growth rate of 2.5% in net plant and equipment per loop from 2007-2010. In contrast only 34% of study areas obtained such a growth rate from 2010-2012. More than 41% of study areas failed to grow in net plant and equipment per loop from 2007-2010; more than 57% of study areas failed to grow from 2010-2012. More than 29% of study areas declined at more than 2.5% for net plant and equipment per loop from 2007-2010; more than 46% declined at a greater rate from 2010-2012. One way to look at Table 20 is as a shift over time. In 2009-2010, 60% of study areas had growth of 2.5% or greater; by 2011-2012, approximately the same number of study areas have only growth greater than zero. In 2009-2010, approximately 52% of study areas had positive growth; by 2011-2012, approximately the same number of study areas had growth of at least -2.5%. Table 20 Percentile Rank for the Rate of Growth of Net Plant per Loop for the 705 Study Areas Rate of Growth 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2.50% 0.563 0.586 0.603 0.671 0.686 0.00% 0.462 0.488 0.526 0.585 0.613-2.50% 0.359 0.393 0.42 0.472 0.517 2007-2010: 0.513 0.414 0.296 2010-2012: 0.657 0.573 0.464 Net financial assets The NECA USF data calculate a measure of net financial assets, or roughly shareholder equity. 28 The greater the net financial assets, presumably the greater capacity a telephone company has to invest and to develop a study area. Table 21 presents information from the NECA database on net financial assets. The net financial assets are roughly the same magnitude a net telecommunications plant in service presented in Table 13. For all 705 study areas, net financial assets range between $8.3 billion and $8.7 billion with mean values between $11.7 million and $12.3 million. Median values are smaller, between $5.7 million and $6.4 million. The largest study areas in the country have net financial assets above $160 million while the smallest ones have net financial assets of less than $100,000. 29 Approximately 80% of study areas have net financial assets between $900,000 and $29.6 million. 28 This is field 220 in the NECA USF database labeled Net plant. This is equal to telecommunications plant and equipment in service plus cash and cash equivalents minus depreciation and amortization and other liabilities of the company. To the extent it captures all of the assets and liabilities of the company, this balance sheet concept is roughly equivalent to shareholder equity. 29 We excluded study areas with negative net financial assets. 21

The mean value for net financial assets increases between 2007 and 2010 from $11.8 million to $12.2 million and then declines from 2010 to 2012 to $11.8 million. This pattern of declines after 2010 also holds for the median value and the 90 th percentile. The 10 th percentile study areas, however, decline consistently from 2007 2012. 2007 2008 2009 2010 2011 2012 mean 11,828,001 11,964,831 12,210,567 12,244,969 12,050,921 11,843,903 Maximum 161,528,942 164,287,558 166,048,736 180,675,079 172,505,854 170,941,463 90th percentile 28,701,528 28,565,260 29,009,434 29,522,054 29,273,945 29,096,767 50th percentile 6,335,112 6,270,164 6,235,993 6,317,352 6,136,124 5,788,235 10th percentile 1,432,097 1,299,420 1,264,960 1,168,817 1,106,380 964,042 Minimum 104,814 83,027 62,320 62,206 67,986 38,258 total 8,338,740,724 8,435,206,004 8,608,449,558 8,632,703,275 8,495,899,450 8,349,951,612 Table 21 Net financial assets for the 705 Study Areas In Table 22, we present the changes in net financial assets from one year to the next. The mean value of changes in net financial assets increases from 2007-2009, declines to $34,000 in 2010, and then declines further in 2011 and 2012 reaching a negative $206,000 in 2012. The median study area declines throughout the period 2007-2012. From the 90 th percentile down to the 10 th percentile, the changes in net financial assets present a consistent picture of decline. 22

Table 22 Changes in Net financial assets for the 705 Study Areas 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 mean 136,830 245,736 34,402 (194,048) (207,018) Maximum 33,290,390 47,775,508 36,186,309 12,666,206 16,121,288 90th percentile 2,012,814 1,906,675 1,802,251 1,421,228 1,351,327 50th percentile (192,260) (193,866) (224,985) (273,829) (256,837) 10th percentile (1,310,791) (1,382,847) (1,473,208) (1,774,077) (1,848,105) Minimum (8,730,484) (14,041,894) (16,735,294) (14,410,349) (17,883,884) total 96,465,280 173,243,554 24,253,717 (136,803,825) (145,947,837) Table 23 presents the percentage change in net financial assets for the period 2007-2012. The mean percentage changes are volatile sensitive to outlier values, such as the maximum percentage change of 2,476% from 2008 to 2009. This one observation accounts for approximately a 3.5% increase in the mean for that year. A pattern still emerges of increasing means for net financial assets from 2007-2010 and declining thereafter. The average increase for the period 2007-2010 is 0.7%; the average decline afterwards is 3.1%. The median percentage changes are not sensitive to outlier values, and the median percentage changes are approximately -5% from 2007-2010 and then increase to above - 7% for the period 2010 2012. The 10 th percentile study areas decline even more rapidly. The ratio of the means from Table 21 is also stable showing increases from 2007-2010 and decreases from 2010-2012. The change in financial assets between 2011 and 2012, whether measured as the mean across 705 study areas, or ratio of the means from Table 5, is between -1.47% and - 1.72%. It is negative in either case. This change is dramatically less than that reported by the FCC s Wireline Competition Bureau for changes in Total Assets of 5.6%. 30 30 FCC Wireline Competition Bureau, Docket 10-90, Universal Service Implementation Progress Report, March 18, 2014., Figure 8. 23