THE MIRROR COMPANIES - PART II J. Paul Groom The auction, for licenses that will compete with the recently privatized fixed wire line properties, is scheduled for November 3rd 1998. An interesting feature of this auction is the method of proposal evaluation. With the B-band mobile cellular licenses there was a two-fold evaluation based upon tariffs and the price to be paid for the concessions. These criteria were weighted by taking 60% of the points given for price proposal, and 40% of the points awarded for the tariff proposal. Price was the only criterion under the Telebrás privatization auction, as tariffs for fixed wire line services are already regulated. However, the "mirror" companies are to be evaluated under a complicated formula evaluating technical proposals as well as price. The purpose of this memorandum is to analyze the equation and then discuss approaches to the auction of these licenses based upon this analysis. The government will auction four "mirror" licenses. There are three regional operating licenses to provide Local and Intra-regional service. The first is known as and the North-East mirror ("Region I"), the second is the Center-South mirror, ("Region II") and the third, the Telesp or São Paulo mirror ("Region III"). The next section will deal with these three licenses. The fourth of these licenses ("Region IV") is for International, Inter regional and Intra-regional long distance, known as the Embratel mirror-company. Regions I, II, and III - Local and Intra-Regional Long Distance Licenses The proposals for the three regional mirror companies will be evaluated using a points system that takes into account price and technical aspects. These criteria will be weighted by multiplying the points awarded for the technical proposal by a factor of 0.7, and the points given for the price proposal by a factor of 0.3. The technical proposal comprises two variables, being the percentages of municipalities within five distinct population bands, that will receive service over a four-year period (1999 - 2002), and the average telephone density offered over the same four-year period. Minimum coverage requirements are 50% of municipalities over 200,000 in population in 1999, 80% by 2000, and 100% by 2001 and 2002. There are 5,448 municipalities in Brazil of which only 107, (1.96%) are over 200,000. Therefore, this is a strangely undemanding minimum requirement. The complete equation is as follows: 0.7 [%MUNij x Factor x {1 + Ln (%DENij / %DMINj)}] + 0.3 [Po / Rp] For analysis purposes we may reduce the equation to three components: 1.% MUNij x Factor. The percentage of Municipalities (% MUN) for each year (j) over five different population bands (i). Anatel?s internet web site includes a list of municipalities, provided as of August 1998 by the Istituto Brasileiro de Geografia e Estatistico - IBGE. The Factor refers to points allocated over the five population bands as seen below: 1999 2000 2001 2002 < 50.000 0,1550600 39% 0,1162950 39% 0,0775290 39% 0,0387650 39% > 50.000 e < 100.000 0,0486810 12% 0,0365110 12% 0,0243410 12% 0,0121710 12% > 100.000 e < 150.000 0,0177370 4% 0,0133030 4% 0,0088690 4% 0,0044340 4% > 150.000 e < 200.000 0,0199400 5% 0,0149550 5% 0,0099700 5% 0,0049850 5% > 200.000 0,1585820 40% 0,1189360 40% 0,0792910 40% 0,0396450 40% Total 0.4 100 % 0.3 100 % 0.2 100 % 0.1 100 % The factor is intended to reflect population distribution within these five categories ( for Regions I and II ), with 39% of points and 44.5% of population in Region I and 42.6% in Region II to municipalities with less than 50,000 population. Region III, due to its smaller size and population density, has a much different population distribution than the other two Regions ( please see Annex I ). The fact that 40% of the points are provided for the cities with over 200,000 population is not meaningful as a competitive factor as 100% of these will certainly have service provided very quickly. 21% of points will be awarded for the provision of service in the three categories between 50,000 and 200,000. The proportion of population in these three categories varies between the regions, from 20.1% in Region I, to 22.3% in Region II, to 23.7% in São Paulo. It should be noted that the total number of points that could be earned in 1999, assuming 100% coverage, is 0.4%. This value declines by 0.1 each year giving an additional weighted benefit for early build-out. 2. {1 + Ln (% DENij / % DMINj)} For the purposes of this equation, density is measured as the number of subscribers per hundred people (pops) in the municipalities where service is to be provided. A minimum density number is provided, namely 0.6% in 1999, 0.8% in 2000, and 1.0% in both 2001 and 2002. The first calculation is to divide the percentage density offered, per year, per population band, by the minimum density. The natural logarithm of the quotient is calculated, and 1 is added. The effect of the natural logarithm is to provide a factor that will diminish the points difference that would have resulted if the simple division of Density Offered/ Minimum Density had been used to calculate the factor. For instance: % DEN/ % DMIN 1 + Ln 1.1666667 1.1541507 1.3333333 1.2876821 1.5000000 1.4054651 1.6666667 1.5108256 The result of this calculation is then multiplied by the result of the first component. This result is then multiplied by the weighting of 0.7. It is not clear why Anatel has included the logarithm in these calculations, as they did not in the points calculations for Region IV. One reason could be that in order for proponents to obtain a differentiating points total they will need more aggressive technical objectives such as more rapid coverage, or greater penetration. A second explanation is that a deflating factor on the technical side of the equation would encourage proponents to accentuate ?price?. 3. Po/ Rp The final calculation is that of dividing the Price Offered by the Reference Price. In the B-band and the privatization auctions there was a minimum price rather than a reference price, so the product of these calculations was always superior to 1.0 (though the points gained in the B-band were then reduced to a factor of 0.6). In this auction reference prices are given solely for the purpose of making this calculation, so it is possible that a proponent could offer a price that is lower than the reference price, and the product of this calculation would be < 1.0. The result of this calculation will then be multiplied by the weighting of 0.3. Competitive Variables Thus, the areas where the battle for these licenses will be fought, will be: Achieving the largest coverage of municipalities under 50,000 in population as early as possible Predicting aggressive, but attainable density levels Price The Municipalities Annex I provides details of the number of municipalities in each population band. It also provides some idea of the magnitude of the build-out necessary for Regions I and II. Region I, has 88.5 million people in 3,030 municipalities, of which 91.3% (2,760) have populations of less than 50,000. These 2,769 municipalities represent only 44.5% of the Region?s population and the average population per municipality under 50,000 is 13,616. Region II has 38.0 million people living in 1,822 municipalities, of which 93.0% (1695) have populations of less than 50,000. The 1,695 municipalities represent 42.6% of the Region?s population and have an average population per municipality under 50,000 of 9,992. By contrast, Region III has a population of 35.3 million in 646 municipalities, of which 83.3% (538) have populations under 50,000. However, these 538 municipalities represent only 19.9% of the Region?s population, and the average population per municipality under 50,000 is 12, 628. It should also be noted that the area to be built out in São Paulo is but a fraction of the enormous land areas that comprise the other two regions, making build-out cheaper and faster and making Region III significantly more attractive. The Points System - Test Case As seen above there are two calculations that are added to provide a final result. One calculation is reduced by a factor of 0.7 (technical), and the other is reduced by a factor of 0.3, (price). This is not the same as saying that 70% of the final points total comes from the technical, and 30% comes from price. As we saw with the B-band bids, an aggressive price proposal can at times outweigh the points gained from other variables. Let us assume that a proponent places a bid for Region I based upon the minimum technical requirements, and a price equal to the reference price. The equation would be as follows: 0.7 [(0.5 x 0.1585820) + (0.8 x 0.1189360) + (1 x 0.0792910) + (1 x 0.0396450)] x {1 + Ln (0.6 / 0.6)} + 0.3 x 1. 0.7 [0.02933758] + 0.3 (1.0) 0.2053631 + 0.3 Final Total = 0.5053631 From this example, it can be seen that the points earned from the minimum technical proposal provide 40.6% of total points, and the price proposal contributes 59.4%. However, there is considerable potential up-side with both the technical and the price proposal, and the proponents will have to decide whether they prefer to earn points through more aggressive build-out, through a higher price quote, or a combination of both. Another way of looking at this is to say that for Region I, an additional R$ 100 million over and above the reference price can be seen as the approximate equivalent of an increase in density from 0.6% to 0.75%, or build-out of 18% of municipalities with a population of less than 50,000. Region IV - Intra-regional, Inter regional and International Long Distance The competition for the nationwide long distance license, the Embratel mirror company, does not have a density component and is, therefore, based upon an equation that takes projected percentage coverage of municipalities multiplied by a factor. The chart below provides the factor per year, per band of population. 1999 2000 2001 2002 < 100.000 0,2037410 51% 0,1528060 51% 0,1018710 51% 0,0509350 51% > 100.000 e < 150.000 0,0177370 4% 0,0133030 4% 0,0088690 4% 0,0044340 4% > 150.000 e < 200.000 0,0199400 5% 0,0149550 5% 0,0099700 5% 0,0049850 5% > 200.000 e < 500.000 0,0527610 13% 0,0395710 13% 0,0263810 13% 0,0131900 13% > 500.000 0,1058210 26% 0,0793650 26% 0,0529090 26% 0,0264560 26% Total 0.4 100 % 0.3 100 % 0.2 100 % 0.1 100 % It will be seen that the population bands are different from those of Regions I, II, and III, with the lowest starting at less than 100,000, and the highest being above 500,000. The points allocation has been changed to match the different population distributions. The most significant weighting change is that the lowest category, less than 100,000, receives 51% of the points and represents 50% of the total population. The relative weighting for early build-out is the same as with Regions I, II and III, being 0.4 in the first year, 1999, and declining by 0.1 each year thereafter. The equation for Region IV is as follows: 0.7 (%MUNij x FACTORij) + 0.3 (Po/Rp). Once again ?i? refers to the population bands, and ?j? to each year from 1999 to 2002. The list of municipalities is the same as that for Regions I, II and III. Minimum coverage requirements are 50% of municipalities over 500,000 in population in 1999, 80% by 2000, and 100% by 2001 and 2002. The Municipalities The winner of the Embratel mirror license will have a minimum coverage requirement of 50% of the municipalities with populations of greater than 500,000 in 1999 (Please see Annex II).This would appear to be even less demanding a coverage requirement than Regions I, II and III. There are 31 municipalities of over 500,000 in population in Brazil, of which 8, or 25.8% are in the State of São Paulo, and 4, (12.9%) in the State of Rio de Janeiro. Thus, with service being provided in São Paulo and Rio de Janeiro only, 80% of the 1999?s requirements have been met. The minimum requirement for 2000 are 80% of municipalities with populations over 500,000, and for 2001 and 2002, 100% of these municipalities must be covered. Competitive Factors 1. The ability of the proponent to service as many municipalities of < 100,000, as early as possible 2. Price It appears that the government will not consider postponing its timetable of November 3rd 1998 for submission of proposals. However, the world economic environment, the current lack of capital resources in the world?s financial markets, and the domestic preoccupations of many USA based telecommunications companies may mean that very few companies and consortia bid for these licenses. Annex I Distribution of Municipalities and Population within Population Bands – Regions I, II, and III. REGION I, ( NORTE LESTE ) Population (in ,000's) < 50 50 e 100 100 e 150 150 e 200 > 200 Mun % Pop Mun % Pop Mun % Pop Mun % Pop Mun % Pop Amazonas 55 37,9% 5 13,6% 1 48,6% Roraima 14 37,5% 1 62,5% Pará 123 45,0% 13 15,2% 3 5,8% 1 2,8% 3 31,2% Amapá 14 23,2% 1 18,5% 1 58,2% Maranhão 200 58,6% 12 15,2% 3 6,7% 2 19,5% Piauí 216 63,8% 3 6,5% 1 4,7% 1 25,1% Ceará 163 45,4% 16 14,9% 1 2,0% 2 5,1% 2 32,6% Rio Grande do Norte 159 55,8% 4 10,3% 2 33,9% Paraíba 215 59,1% 5 10,2% 1 3,3% 2 27,5% Pernambuco 147 39,1% 18 15,1% 4 6,5% 6 39,3% Alagoas 95 56,0% 4 8,9% 1 6,6% 1 28,5% Sergipe 69 50,4% 4 16,1% 1 7,3% 1 26,1% Bahia 378 52,7% 25 13,1% 5 4,7% 3 4,3% 4 25,3% Espirito Santo 65 36,8% 4 9,6% 2 7,3% 1 5,3% 4 41,0% Minas Gerais 797 47,5% 35 14,7% 8 5,3% 3 3,0% 10 29,3% Rio de Janeiro 59 8,8% 12 6,6% 4 3,5% 5 6,2% 10 74,9% Total 2.769 44,5% 161 13,5% 33 4,4% 17 4,4% 50 33,2% Annex I – ( continued ) REGION II, ( CENTRO SUL ) Population (in ,000's) < 50 50 e 100 100 e 150 150 e 200 > 200 Mun % Pop Mun % Pop Mun % Pop Mun % Pop Mun % Pop Mato Grosso 120 57,1% 3 8,3% 1 6,5% 2 28,1% Mato Grosso do Sul 72 48,7% 3 11,5% 1 8,1% 1 31,8% Rio Grande do Sul 427 39,1% 23 16,1% 5 5,8% 4 7,1% 8 31,9% Santa Catarina 275 52,9% 10 13,7% 3 8,4% 1 3,0% 4 21,9% Paraná 370 45,2% 18 14,3% 2 2,4% 3 5,6% 6 32,4% Tocantins 136 73,7% 1 6,2% 2 20,1% Acre 20 39,5% 1 11,8% 5 48,6% Rondônia 48 57,3% 3 18,8% 1 23,9% Goiás 227 45,8% 10 15,3% 2 4,7% 3 34,2% D.F. 1 100,0% Total 1.695 42,6% 72 12,4% 14 4,4% 10 3,3% 31 37,3% REGION III, ( SÃO PAULO ) Population (in ,000's) < 50 50 e 100 100 e 150 150 e 200 > 200 Mun % Pop Mun % Pop Mun % Pop Mun % Pop Mun % Pop São Paulo 538 19,9% 47 9,6% 21 7,0% 14 7,1% 26 56,0% Annex II Distribution of Municipalities and Population within Population Bands – Region IV REGION IV - NATIONWIDE Population (in ,000's) < 100 100 e 150 150 e 200 250 e 500 > 500 Mun % Pop Mun % Pop Mun % Pop Mun % Pop Mun % Pop Amazonas 60 51.5% 1 48.6% Roraima 14 37.5% 1 62.5% 0.0% Para 136 60.2% 3 5.8% 1 2.8% 2 10.8% 1 20.4% Amapá 15 41.7% 1 58.2% 0.0% Maranhão 212 73.8% 3 6.7% 1 4.2% 1 15.3% Piauí 219 70.3% 1 4.7% 1 25.1% Ceara 179 60.3% 1 2.0% 2 5.1% 1 3.3% 1 29.3% Rio Grande do Norte 163 66.1% 1 8.1% 1 25.8% Paraíba 220 69.3% 1 3.3% 1 10.4% 1 17.1% Pernambuco 165 54.2% 4 6.5% 4 13.8% 2 25.5% Alagoas 99 64.9% 1 6.6% 1 28.5% Sergipe 73 66.5% 1 7.3% 1 26.1% 0.0% Bahia 403 65.8% 5 4.7% 3 4.3% 3 7.6% 1 17.7% Espirito Santo 69 46.4% 2 7.3% 1 5.3% 4 41.0% 0.0% Minas Gerais 832 62.2% 8 5.3% 3 3.0% 8 13.9% 2 15.4% Rio de Janeiro 71 15.4% 4 3.5% 5 6.2% 6 16.2% 4 58.7% Mato Grosso 123 65.4% 1 6.5% 2 28.1% Mato Grosso do Sul 75 60.2% 1 8.1% 1 31.8% Rio Grande do Sul 450 55.2% 5 5.8% 4 7.1% 7 18.6% 1 13.3% Santa Catarina 285 66.6% 3 8.4% 1 3.0% 3 18.6% 1 3.3% Paraná 388 59.5% 2 2.4% 3 5.6% 5 15.7% 1 16.7% Tocantins 137 79.9% 2 20.1% Acre 21 51.3% 1 48.6% Rondônia 51 76.1% 1 23.9% Goiás 237 61.1% 2 4.7% 2 12.3% 1 21.9% D.F. 1 100.0% São Paulo 585 29.5% 21 7.0% 14 7.1% 18 16.2% 8 39.8% Total 5.282 50.0% 68 5.0% 41 4.5% 72 13.5% 31 27.0%

volta