Document(s) 19 of 21

APPENDIX 1

Background and Methodology: Survey of Telecom Use at the Bottom of the Pyramid

The findings reported in this book are based on a subset of findings of a larger knowledge, attitude and practice study of the telecom usage patterns and behaviors of a sample of 'financially constrained' 'users' in 11 localities in India and Sri Lanka, conducted by LIRNE asia. Face-to-face interviews were conducted in both countries with a total of 3,199 respondents in April and May 2005. Seven localities were surveyed in India and four in Sri Lanka (Table A1). With the exception of Colombo (Sri Lanka) and Mumbai (India), interviewees were spread across urban and rural areas of each locality. The questionnaire was translated into, and conducted in, five local languages (Hindi, Malayalam, Oriya, Sinhala, and Tamil).

Table A1.1
Distribution of Respondents among Localities Studied in the Survey of Telecom Use at the Bottom of the Pyramid

For the purposes of this study, the 'financially constrained' were defined by two parameters; first, those with household income levels of approximately USD 100;¹ second, socio-economic levels. In Sri Lanka those belonging to socio-economic classification² (SEC) groups 'B', 'C', 'D', or 'E' were included in the sample. In the Indian sample, a different, but comparable socio-economic classification was used. Socio-economic classification of the 'financially constrained' in India according to the natural distribution of population is divided among urban and rural settings, each consisting of different SEC groups. In urban India the 'financially constrained' can be classified as SEC 'B', 'C', 'D', and 'E', while rural 'financially constrained' in India can be classified as 'R1', 'R2', 'R3', and 'R4' based on the profession and type of dwelling of the chief wage earner (pucca and kuchha house). In this study, this division was followed for the socio-economic classification of Indian users.

Respondents over 18 years old were chosen from within selected households³ based on Kish sampling techniques⁴ to ensure random sampling as well as adequate representation of gender and age groups as in their actually existing ratios.

India and Sri Lanka are located in South Asia, the largest concentration of poor people in the world. Both countries have experienced rapid telecom growth within the past five years. In addition, India and Sri Lanka have differing mobile termination regimes: India is a Calling Party Pays (CPP) environment (from 2003), similar to the regime in fixed where the service of receiving a call is bundled together with call origination, which is charged; Sri Lanka is a Receiving Party Pays (RPP) environment, where one has to pay for both origination and reception, though many consumers now enjoy significant quantities of free incoming minutes under various packages. It was hoped that this study might bring out the differences, if any, in telecom use among the 'financially constrained' in the two environments. The seven different localities in India and the four in Sri Lanka were selected, not to represent the two countries, but to capture the diversity within the two countries, taking snapshots of 11 very different markets, in terms of telecom access, economy, population, and geography. For this purpose, the 'Indian' sample was further divided into two 'regions' for some of the analysis: 'Northern' India (Dehradun, Gorakhpur, and Neemuch) and 'Southern' India (Cuttack, Kasargod, Mumbai, and Sivaganga). The rationale for grouping the locations was the broad similarity in the socio-economic qualities of the locations. This was done in an attempt to preserve some of the diversity of the locations, as well as to split the sample more evenly for comparison.

Limitations

The findings from this study are not representative of India and Sri Lanka as wholes. A true representation could only be obtained through pure random sampling according to the natural distribution of the population in the countries, rather than purposive sampling of the localities which was undertaken. However, the individual locality samples are representative of the 'financially constrained', as defined by this study.

Respondents were asked to indicate their monthly income for the purpose of analysis, including income from all sources, which means that the income reported would have been that for the household. While respondents were asked to consider income from all sources, it is still plausible that the income group indicated does not reflect true income levels; irregularities in remittances, which can account for substantial portions of income in developing country households, where large sums of money are received at irregular intervals for special occasions or emergency situations, could result in some income being unreported. Furthermore, such transfers may not even be considered as 'income' per se by the respondents.

For the kind of information that this survey sought to elicit, a questionnaire containing many more open-ended questions would have been optimal. However, given the size of the sample as well as the depth of the questionnaire, this was not practical. For this reason, the questions were closed-ended, but respondents were given many nonexclusive options to choose from.

The survey asked respondents about their calling patterns, in terms of average number of calls made and received per month, etc., to what destinations and for what purpose. It is recognized that the accuracy of this information is problematic because it is based on recollection. Thus the data obtained is only an indication of and not necessarily an accurate representation of individual calling patterns. Real calling patterns can only be obtained from billing records; this was not done in this study for privacy reasons. In any case, the option of analyzing billing records exists only for a small percentage portion of the sample, the 26 percent of fixed phones owners and the 2 percent of post-paid mobile owners.

There is also over representation of 'unemployed' persons and 'housewives' in the sample. This could not have been avoided unless quota sampling was adopted by occupation categories.

One significant weakness of this study is that it does not study the financially 'unconstrained.' Sound conclusions about the behavior of the financially constrained can only be made if the financially 'unconstrained' are studied in comparison, that is, through a sample which also covers the SEC A's, and those with monthly incomes household over US dollars (USD) 100 per month. The study also does not look at non-users amongst the financially constrained, and how their non-use is associated with financial constraints. Furthermore, it is not possible to say whether behavioral patterns identified in this study are also relevant to the financially constrained in more developed markets, or are unique to the financially constrained of South Asia alone, without studying comparable data for those markets as well.

This research has served as a pilot from which LIRNEasia has increased its understanding of telecom use by the financially constrained, as well as identified areas that can be improved upon in the research, which has helped shape LIRNEasia's 2006–2007 research in this area to better understand this use, in a larger group of countries in South and Southeast Asia.

APPENDIX 2

Supporting Information for Chapter 8

Table A2.1
Rural Telecoms Service (RTS) Maximum Tariffs and Default Interconnection Rates in Nepal

Source: Annex 17 to RFA for RTS.

Notes: a All tariffs are in rates per minute, unless otherwise indicated.

^b Paid to NTC or other operator, unless different rate is mutually agreed.

^c Payable by NTC or other operator, unless different rate is mutually agreed.

^d Tariffs to be subject to price cap indexing after 2004 in accordance with Tariff Guidelines.

^e Termination charges prescribed in Guidelines on Interconnection (GI).

N/A means 'Not Applicable'

Table A2.2
Technology Cost Guidelines of Nepali Least Cost Subsidy Auctions

Source: The World Bank Discussion Paper, No. 432.

APPENDIX 3

Supporting Information for Chapter 9

Bidding Process for the Provision of Rural Household Direct Exchange Lines (RDELs) in High Cost Specified Short Distance Calling Areas (SDCAs) in India

1. The bidding process has been structured as 'Multi-layered Informed Descending Bidding Process'. The pre-qualification bid along with EMBG and the first Financial Bid will be submitted by the bidders in separate sealed covers. The first Financial Bids of those who qualify shall be opened and shall be made public in the presence of bidders' authorized representatives, carrying such authority/document with them.
   
   
2. The bidders can offer less than or equal to the benchmark for the front-loaded subsidy. Offers higher than the benchmark for the front-loaded subsidy, or where all the specified SDCAs in an SSA have not been bid for, shall not be treated as a valid bid.
   
   
3. The lowest valid offer for an SSA will determine the reserve price for the front loaded subsidy for each SDCA in an SSA for the second round. If there are more than four pre-qualified bidders for an SSA, the highest bidder will be dropped and others will qualify for the second round of financial bidding. If pre-qualified bidders are four or less, all will qualify for second round of financial bidding. If there is only one bidder with a valid bid, then that bidder will be declared as the successful bidder and there shall be no further round of bidding.
   
   
4. The bidders, except those last dropped, shall submit a Second Financial Bid for the second round. The bid amount in the second round will have to be equal to or lower than the 'Reserve Price' for each of the SDCA in the SSA and the bidder(s) quoting above the Reserve Price shall be disqualified for further round of bidding.
   
   
5. All the bidders who have quoted less than or equal to the 'Reserve Price' for the front loaded subsidy for each of the SDCAs in an SSA in the second round of financial bidding will be short-listed. If there are three or more such short-listed bidders in an SSA, except the highest bidder, the rest will qualify for the third round of financial bidding. In case there are two bidders, both will qualify for the third round of financial bidding. In case there is only one short-listed bidder, then that bidder will be declared the successful bidder and there will be no further round of bidding.
   
   
6. The lowest valid offer for an SSA will determine the Reserve Price for the front loaded subsidy for each SDCA within an SSA for the third round. The bidders, who qualify for bidding for the third and final round, shall be required to submit the third and final bid for all specified SDCAs within an SSA for which they have qualified in the previous round. The bidder of the final round of financial bidding with the lowest offer for an SSA will be declared successful for signing of the Agreement.
   
   
7. In case of a tie in the conclusive round of bidding, the bidder who quoted lower amount in the previous round for an SSA shall be declared successful for signing of the Agreement.
   
   
8. In the event of a tie in the previous bidding for selection, the process of bidding will continue until a successful bidder emerges on comparison.
   
   
9. The signing of Agreement as a result of bidding process will not be treated and taken as grant of fresh License under the Indian Telegraph Act, 1885.
   
   
10. If the Agreement is not signed with the successful bidder for whatsoever reason, the Administrator may decide to continue with the present bidding process till it is completed and the Agreement is signed with the successful bidder.
    
    
11. In every successive round of bidding, the bidders can only lower or retain the front loaded subsidy of their last bid. Withdrawal or backing out of their last quoted front loaded subsidy will attract forfeiture of EMBG.

Table A3.1
License Fees of Various Kinds of Telecom Operators in India

NOTES

¹ Approximately INR 5,000 in India and LKR 10,000 in Sri Lanka at the time.

² A standard classification, based on occupation and education level of the chief wage earner.

³. A maximum of five households were selected starting from one 'starting' household that was randomly selected from the electoral list.

⁴ The Kish Grid is a random sampling technique to select one respondent from many eligible respondents in a household. In this case, names, gender, and ages of all household members using phones (in the preceding three months) were recorded (in descending order of age). Based on the number of eligible respondents in a household and the household contact number (nth interview of each starting point), a random number sheet was used to select one of the many eligible respondents. This ensures that respondents selected are not skewed to any gender or age, but are reflective of reality.

Document(s) 19 of 21