Abstract
Nepali information and communication technology (ICT) policies, informed by dominant international discourses on internet and ICT for development, have attributed a ‘self-evident’ positive impact to economic growth. Such a conception is inconclusive and hides the underlying socioeconomic inequalities and the barriers they oppose to universal connectivity ambitions. The self-evident discourse overlooks the importance of adjacent infrastructures like electricity, critical for efficient functioning of a ‘digital society’. History has shown Nepal’s eagerness to buy the irrational exuberance of the dot-com era. It also shows policy targets were not based on local evidence and experiences. While the question about when the next bubble will emerge and burst is debatable, the techno-deterministic arguments need to be checked in light of evidence. We offer insights from import trajectories since the drafting of the first IT policy. The proliferation of mobile phones is explored in light of data consumption (i.e., usage) and access during the deadly earthquake in 2015. The evidence clearly highlights the mismatch between connectivity ambitions and socioeconomic realities in a stratified information society. We contend that the self-evident visions of connectivity need to be checked empirically. We call for internet and ICT policies that have a nuanced conceptualisation of access and its gradation.
Introduction
ICT has received significant attention since it was posited as a key instrument for achieving the millennium development goals. Thereafter, many international organisations began to consider the benefits of internet as self-evident, which is exemplified by assertions like ‘information and communication technologies (ICTs) have proven to be a tremendous accelerator of economic and social progress’ (UNCTAD, 2009) and countries failing to join in on the broadband revolution ‘will lose the opportunity to reap the economic and social benefits’ (Budde, 2011). More recently, increasing access to ICT and internet is set as a direct target of goal 9 in the sustainable development goals and is a key instrument in most sustainable development programmes. Agencies such as International Telecommunication Union (ITU) point that economic development, social inclusion and environmental protection is impossible without the involvement of ICT, and lobby for its crucial role in achieving all 17 goals. Such optimism is at the heart of internet and ICT policies of Nepal.
Academic research on the impact of ICTs is mixed and inconclusive. Numerous studies point out that ICT in isolation may not lead to development and can even multiply the existing inequalities (e.g., Friederici et al., 2016; Steyn, 2016; Tyson & Spence, 2017; van Deursen & van Dijk, 2014). Such studies find transformative impacts of ICTs are variegated, contingent and inconclusive. They ask for justification of the self-evident discourses. Policies in low-income countries like Nepal are rooted in self-evident discourses of select international organisations. The implicit assumptions and preconditions that associate connectivity and growth lack evidence. For instance, Nepali policies do not mention that ICT’s potential is constrained by access, affordability and quality of service (Regmi, 2017). There is also no discussion on whether favourable business environment, strong human capital and good governance are sufficient to assert self-evident claims. The ground realities show that connectivity is disproportionately limited in Nepal, a fact often understated by the simplistic equating of access with subscription numbers (Pandey & Raj, 2016; Pandey & Regmi, 2018).
This article contends that Nepali internet and ICT policies are informed mostly by a self-evident vision of connectivity despite highly uneven economic impact of internet connectivity across geographies and socioeconomic strata. Such a conception hides the underlying socioeconomic inequalities and the barriers they pose to internet adoption and benefits acquired from its use. It also overlooks the centrality of supportive physical infrastructures like roads and electricity for efficient functioning of a ‘digital society’. It is left as an exercise post-connectivity, something that can be fixed.
The article is organised as follows. It starts with a short review of Nepali ICT related plans and policies since the 1990s. The analysis reveals how the discourses have maintained an overly optimistic and simplistic view on the impacts of ICT and connectivity. In the next section, we show how Nepal was quick to buy into the irrational exuberance surrounding the internet between late nineties and early 2000s. We also show how it did not last very long coinciding with the dot-com bubble. We then provide evidence to reveal the underlying inequalities before 2010 (the computer and internet boom) and after 2015 (mobile telecommunications boom). We also make a basic estimation of inequality in mobile data consumption. Next, we discuss how the policy visions have failed to address the underlying inequality and state the importance of adjacent infrastructures. Finally, we end the article with the conclusion. We argue for more grounded internet policies that cautiously assess the discourses about the impacts of ICT and forego techno-determinism and self-evident premises.
The ‘Capacity Centric’ Policies
Nepali internet and ICT policies are informed by dominant narratives that claim favourable impact of ICT on development, specifically economic growth. Simplistic access indicators, such as subscription numbers or infrastructure spending, are taken as an alias for technology adoption. Most large-scale surveys such as the national census and living standard survey do not ask much beyond simple questions such as ‘Do you have internet at home?’ and ‘How often do you use the internet?’. There is some interest in ownership of the devices but no consideration is given to how people use them and what they get from this use. Access is not viewed as a technosocial system. Technological determinism dictates access to be understood in terms of capacity and reach, putting the people as non-relevant for societal outcomes.
The importance of digital technologies was officially addressed during the late nineties in Nepal, following a large expansion of the telecommunication infrastructure. Computerisation was seen as a solution to economic ills and decline in productivity. The use of computers to process national census of 1971 had reduced the time to produce the report by years. Subsequently, the eighth five-year plan (1992) talks about procurement of computers and computerisation of data on forestry, land ownership records and radio frequency index (Government of Nepal, 1992). Increasing tele-density was also a way to increase dial-up internet connectivity on the long run. From public institutions point of view, imports and policy efforts in the nineties were largely indicative of a move towards computerisation.
The seeds of a ‘knowledge-society’ were sown in the 1990s, when the ninth five-year plan addressed the importance of internet infrastructure (Government of Nepal, 1997). This period fits snuggly with the period of ‘irrational exuberance’ of dot-com funding, known as dot-com bubble (roughly 1995–2001), when the optimism of the internet growth was ‘riding bubbles’. The next section shows how this quickly translated into escalated level of technology import. In Nepal, developments at the regulatory and policy front showed the knock-on effect of international optimism. Telecommunications act (1997) and later telecommunication policy (2004) are considered instrumental in the liberalisation of the telecommunication sector and in promoting private investment. Ministry of Science and Technology (MoST) was created in 1996 with a role to create conducive environment for effective application of science and technology including the information technology sector. The Nepal Telecommunications Authority (NTA) was established in 1998 as an autonomous regulatory body. Subsequently the 2057 (2000 A.D.) information technology policy of Nepal envisaged a transformed knowledge-based society. High-Level Commission for Information Technology (HLCIT) was later formed in 2003 under the chairmanship of the Prime Minister to provide a strategic direction to its information technology policy.
Though the initial focus was on broadband, recent internet policies in Nepal talk about universal access to services due to the proliferation of mobile phones. Mobile phone plays a central role to ensure by 2020 (a) access to 80 per cent of citizen facing services; and (b) internet access to the entire population (Government of Nepal, 2015b, p. 3). 1 In contrast to the slow adoption of computers, mobile phones spread like wildfire after 2010 in relation to the development of 3G and General Packet Radio Services (GPRS) technologies. Akin to increasing access to computers and trained computer users, ICT policy of 2015 talks about increasing internet access and skilled internet users. Note that the shift of focus from computer to the broader ICT coincides with a global push for these technologies. Arguments for computerisation, abundant in the 1990s, shifted towards ICTs from the mid-2000s. Global discourse is now dominated by telecommunication technologies associated with mobile telephony.
The sequences of policies spanning over more than two decades share an important postulation. The ambition is to deliver overarching policy goals touching all sectors of the economy at once. Similar to what Friederici et al. (2016) observe in the context of Africa, Nepali policies also claim that universal connectivity translates to economic growth and inclusive development. Similarly, how the technologies operate in an unequal socioeconomic environment is missing. Table 1 shows the claims and barriers to ambitions as depicted in the key plan and policy documents. Links between connectivity and economic growth are a consistent connection. Barriers are mostly reflective of the rural and urban gap. Relatively simple fixes such as training to improve digital skills and regulatory mechanisms to cater future services and problems that arise from it are proposed. Although electricity shortage is acknowledged, there is neither an energy audit of the present ICT sector or projections of the proposed ‘Digital Nepal’ ambitions. It is to be noted that shortage of electricity does not limit the scale of ambitions even in the short run where clearly electricity is not in surplus and has to be imported from India.
We have already stated that the claims about ICT macro-level economic benefits are relying on inconclusive evidence (Friederici et al., 2016): there are studies confirming both negative and positive outcomes. The context is also missing in the policy narratives. Positive effects are mostly observed in high-income countries and in urban locations. How the transformations translate to low-income country like Nepal is unclear.
The Presumed Impacts of Connectivity and Its Barriers in Nepal’s Information Technology Related Plans and Policies
Some studies do not find a relationship between ICT and economic growth even for high economies. Ishida (2015) estimates that there is no long-run relation between investments in ICT and economic growth in Japan. At the micro-level, theoretical models used in ICT and development studies are often too simplistic with strong assumptions that do not meet the ground realities. For instance, the claim that mobile phones were instrumental in reducing price dispersion and eventually benefiting fishermen in Kerala has been strongly criticised (Steyn, 2016).
Moreover, other ethnographic studies suggest mobile phones and ICT in general have no relevance for economic activities in the rural areas of developing countries (Arora, 2016; Burrell & Oreglia, 2015). Regarding the adverse effects, Tyson and Spence (2017) argue internet is a skill-biased technology, which can widen the prevailing inequalities in society. There is a long list of studies that show little or no impact of ICT on development, be it economic or of other types, specifically in the case of low-income countries. There is a cause for caution and a call for evidence.
Access defined as reach, which is taken as an alias for connectivity, is the only indicator of success in the information technology policies. Broadband policy 2015 had targeted to provide internet access to 45 per cent of the households within 2018. Such targets tend to be global averages in the absence of evidence and influenced by international organisations. An oft-cited claim of the World Bank is that a 10 per cent increase in ICT infrastructure investments causes GDP to grow by 1.38 percentage points (Qiang et al., 2009). Koutroumpis (2009) estimated a minimum threshold of 20 per cent household internet penetration to have significant effect on economic growth.
Reflections from Import Statistics
Import statistics is a good way to gauge activity of an import-based sector and its spending on infrastructure. Throughout the mid-nineties to early 2000s Nepal was a leader in technology imports in South Asia. The large bump in the late nineties to the early 2000s is a reflection of large spending on infrastructure. No other countries in the region comes close to Nepal. This is despite Nepal not having the workforce like India or an industrial set-up like Singapore to capitalise on global connectivity. Unsurprisingly there is a sharp decline of ICT imports after 2000. We find the bump in ICT imports aligns with optimism around the telecommunication industry and the internet (dot-com). This section examines these distinctive patterns.
This section uses imports data classified using SITC (Standard International Trade Classification) provided by the Observatory of Economic Complexity (Simoes and Hidalgo, 2011). Commodities under two main headings have been studied: (a) electronic data processing (EDP) equipment (SITC code 752) and (b) telecommunication equipment (SITC code 764). EDP equipment includes personal computers, workstations, hard drives and peripherals such as keyboards, mouse, printers and monitors. Telecommunication equipment include telephone sets, transmission equipment, television and radio transmitters, switching equipment and other telecommunication equipment, parts and accessories. We have used the ICT deflator from the United States Bureau of Labour Statistics (BLS) to estimate the real values on imports. The deflator has been adjusted for domestic inflation using the exchange rate of each country (Daveri, 2001). We have used United States numbers because statistical bodies in Nepal and other South Asian countries do not maintain a deflator for ICT goods. For our objective, the adjusted price does not have to be exacted. We look at trends but make no forecasts on import figures. In few places, we have left the values at current price when it does not affect the findings. We also use diffusion and consumption statistics provided by NTA to tie in the trends with present situation. All data used in this article are publicly available.
The exuberance of the ICT sector is reflected in the large expenditure on EDP equipment. In Figure 1, the unmissable bump for EDP imports fits within the period of dot-com bubble from 1995 to 2000. The level of imports in the dot-com bubble period has only been matched after 2010. For telecommunication goods, the point of rapid increase in imports coincides with the introduction of 3G and GPRS services in 2007. The curve for telecom imports has a much sharper ascend from 2010 onwards. The usually attributed reason for this sharp ascend is the presence of strong private competitors to then state-owned, now public telecommunication company, Nepal Telecom. 2
We also see an increase in Nepali telecommunication imports during the 2002–2006 period. During this time, optical fibre along the East–West highway covering 904 km was laid at the cost of twenty million USD (Pangeni, 2016). Soon after another 115 km (North–South) optical information highway was laid at about four million USD. These fibre links would act as a backbone for the mobile services. To expand its subscriber base Nepal Telecom invested thirty million USD to double its installed base (Whalley, 2006). This aside, major growth in the telecommunication sector of Nepal is attributed to initiatives during the late nineties. Specifically, loans amounting to around 100 million USD from the World Bank are credited with the transformation of the telecommunication infrastructure (Goodman et al., 2000). We can see a distinctive rise in Telecom imports around that time in Figure 2(b). This explains the doubling of the 1997 landline subscribers in just two years. But the waiting list for a telephone connection in Nepal was long, six years in 1999 (Goodman et al., 2000). Non-receding waiting list was a characteristic of the region at the time.
The telecommunication sector was unable to satisfy demand for the landline phone in the following years. It took six years to add 1 per cent (of the population) landline subscribers from 1999. The rural telecommunication development fund (RTDF) that could have helped in this venture was not utilised until 2016. The fund that started around 1999/2000 for the development of telecommunication services in rural Nepal, until recently, had accumulated up to 125 million USD with an annual contribution of twenty-four million from the licensees. Due to inactivity, the reach of telecommunication services is highly disproportionate across rural/urban and geographical terrains. Lately, during the last three years, NTA has allocated the RTDF to connect the districts of Nepal through an optical fibre network.
From Figure 2, it is clear that Nepal was first in the SAARC region to latch on to the telecom boom in the late 1990s. It was also the first to buy into the dot-com boom at a similar time. There are distinct bumps during the late 1990s and early 2000s that are the highest in the region. The exuberance however did not last long and it coincides with the dot-com bubble and telecom bust. Other countries in the SAARC region also saw a rise in imports during the boom period, but not at such dramatic levels. While most countries returned to the pre-boom levels, Nepal had a much deeper fall.
One reason for the sharp fall of imports after 2000 in Nepal can be the political instability in the country. Nepal was going through a period of Maoist insurgency at the time. Royal massacre also took place in 2001. Some commentators have pointed rather at the shift in priorities for the reason behind under resourced implementing bodies. Burton (2003) observes this shift in combination with decline in global optimism and conflict between related policies in other ministries as reasons for the lack of progress. The conflict can also be understood from the shift in policy. IT policy, which was under the MoST in 2001, was superseded by ICT policy, which is now under the jurisdiction of Ministry of Communication and Information Technology since 2015. It can be argued this move is an effect of the success of mobile telecommunications and mobile internet, becoming synonymous with internet in Nepal. Large investments in the telecom sector are seen around 2005–2008 in preparation for the rise in mobile telecommunications uptake.
Comparison of Human and ICT Development in the South Asian Countries Considered
A look at some development indicators suggest that import followed enthusiasm around technologies that could get rid of economic and development ills. In Table 2, except for Sri Lanka that has high values for the development indicators, Nepal ranks almost equal with India, Bangladesh and Pakistan. Despite this similarity, imports in Nepal were significantly higher in the region. The potential had trumped reality. After 2008, EDP and telecom imports for Nepal are under two and three per cent of total imports, which are comparable to the imports of other countries. It can be argued that telecoms are cautious and expansion in infrastructure is not as rapid. In Nepal, two telecoms share the market equally and there is no urgency. This also explains the poor quality-of-service that is regularly making rounds in daily newspapers and surveys.
The Perennial Inequality
We have so far made a case for how activities in the policy regime and investments in the ICT sector were riding on enthusiasm around the telecommunications industry and the potential of the internet. We now explore the underlying realities, the necessary pre-conditions, which were sidelined by such optimism surrounding the technologies. We split the timeline into before 2010 and after 2015 periods. Major internet and ICT policies were drafted in these windows. We specifically focus on the latter as the diffusion of mobile technologies has seen an unprecedented growth in reach and ownership vastly surpassing previous digital technologies. We also provide data on consumption and use that gives a richer insight into access, beyond subscription numbers.
Inequalities in the Early period
Fifty-six per cent of Nepal’s village development committees had no telecommunications service in 2000 (Intven et al., 2004). Kathmandu valley alone accounted for two-thirds of the total telephone sets (Goodman et al., 2000). Electricity and literacy were seen as necessary pre-conditions to achieve the access targets. These were crucial to realise the development transformation that dial-up internet and computers were going to bring.
Access to electricity and levels of literacy were disproportionately distributed between rural and urban areas. Only 14 per cent of all households and less than 10 per cent of all rural households had electricity access around 1996 (Nepal South Asia Center, 1998). Even the information technology indicators were misleading, as access statistics did not describe the actual consumption of technology. In rural Nepal, though there were telephone booths, the terrain seriously had limited access (Pigato, 2001). Nepal also did not have the niche market and a software workforce like India to justify the policy targets. Nor it had an export-oriented industry like Singapore. In that context, the first IT policy (Government of Nepal, 2000), targeting ‘to place Nepal on the global map of information technology within the next five years’, was ambitious and unrealistic. The agenda was set on global exuberance given the barriers it needed to remove and lack of provision to exploit the new developments. Table 3 runs down some important indices around 2000 that were considered relevant by the plans and policies. The table also explicates the urban–rural gap that was central to the universal service ambitions.
Important Access Indicators During Drafting of the First Information Technology Policy
Similar rural–urban disparity and uneven geographical diffusion of ICT devices and services is visible in the 2011 nation census. At the district level, cluster analysis by Pandey & Regmi (2018) show access to ICT facilities has a clear geographical pattern. They find there is almost no fixed-internet penetration in thirty-two out of the seventy-five districts. Location advantage (proximity to major cities) and favourable terrain are strong indicators of ICT infrastructure and return-on-investment for the service providers. Similarly, locations with the best terrain, higher population density and higher income generally have better 3G site coverage (Pandey & Regmi, 2018). However, not all favourable locations see the number of users and volume of traffic that correlates with these indicators. It would be incorrect to assume a causal relationship between access and data consumption.
Inequality in electricity usage also persisted at the household level. The 2011 Nepal Living Standards Survey (CBS, 2011) shows top 25 per cent of household spent NPR 7,158 (approx. USD 71) annually in 2011. But the lower 75 per cent spent only NPR 657 (approx. USD 0.6). Electricity expenditure for the lower 50 per cent of the households is literally zero in the survey, as they did not have a household electricity connection. This again demonstrates that access does not necessarily translate into use.
Inequalities after 2015
Despite the ubiquity of mobile phones, data consumption is very low in Nepal. Table 2 shows that on average, a data subscriber to Nepal Telecom and Ncell, the two leading telecommunication companies in Nepal, spends only 0.43 and 0.66 USD per month on data. In view of data consumption, considering the data tariffs during 2015/16, an individual Nepal Telecom subscriber used 84 MB of data on average, and Ncell subscriber used 43 MB more. This amount of data translates to watching just an hour of YouTube video on the lowest quality setting (240p) over a thirty-day period. After the introduction of 4G services, operators are providing data at much cheaper rates, such as NPR 100 (approx. USD 1) for 1 GB of data, but with stringent expiration constraints.
The statistics presented in Table 4 do not make sense at face value. They only make sense if we consider the phenomena that a large proportion of subscribers consume small volume of data, whereas a small proportion of heavy users consume most of the data. A large majority of subscribers are somewhere in between. As the cost of data has come down, the heavy users are using a lot more data than other groups.
Average Earnings from Telecommunication Services
The heavy usage is also limited to certain locations of the major urban hubs. An analysis of the tweets during the Gorkha earthquake of 2015 showed 60 per cent of total tweets originate from the Bagmati zone, which contains the three major urban centres of Nepal, viz. Kathmandu, Bhaktapur and Lalitpur (Pandey and Regmi, 2018). Forty-one per cent of total tweets originate from a single city (Kathmandu) in 2015 and 2017. All top ten places in tweet volume are located in the Kathmandu valley. Sixteen districts (out of seventy-five) fail to register any Twitter activity. Similar tendencies are also found in volume of Google search queries.
The observation that only a small proportion of mobile users are responsible for large volume of data is a global phenomenon. In 2010, 3 per cent mobile users of a large 3G network operator in the United States contributed to half of the mobile data traffic (Jin et al., 2012, pp. 7–12). A more recent study on the mobile traffic of a telecom operator in southern China showed 1 per cent of the users consumed 9.98 TB out of the total 11.39 TB data in a week (Yang et al., 2015). Apart from these academic studies, reports by multinational companies also provide similar statistics. The report by Amdocs (2015), for instance, revealed that 10 per cent of users consume 80 per cent of network data in major cities of the world. In advanced economies where proliferation of smartphones might have reached the ubiquitous level, inequality is now observable in data consumption.
Household surveys also back the right-skewed distribution assertion mentioned above. Average monthly income of the richest 20 per cent Nepali household is 518 USD (NPR 53,578) in 2015 (Nepal Rastra Bank, 2015, p. 26). For the poorest 20 per cent, average monthly income is 177 USD (NPR 18,338). Regarding communication expenses, the richest 20 per cent of households spend USD 16.7 (NPR 1,736), whereas the poorest 20 per cent spend USD 3.4 (NPR 356). Average Nepali household size is 4.88. Thus, communication expenses per person for the richest and poorest 20 per cent of the households are USD 36.2 and USD 3.4, respectively.
It is evident from Table 4 that communication expenses of the lower 20 per cent is almost equal to the average voice expenses per subscriber. The upper 20 per cent household has a surplus of about USD 13 per person for communication expenses. As mobile phone is the de facto medium of communication, the surplus largely corresponds to mobile data usage. This also supports the assertion that mobile data consumption in Nepal also has a highly skewed distribution.
Estimated Average Data Expenses in 2016 for Various Expenditure Groups
To get an idea of the picture of data usage inequality in Nepal, we have made a preliminary estimation of the mobile data expenses for various expenditure groups in Table 5. 9 Given the average data consumption for NTC and Ncell users in 2015/16, we made an estimation assuming data consumption is exponentially distributed across the mobile internet users. The estimation shows a stark difference in data consumption between the bottom 90 per cent and upper 10 per cent of the users. The average estimated data consumption for the bottom 10 per cent of NTC data users is 0.32 USD per month, whereas the top 10 per cent average estimated expense is 1.42 USD per month. Ncell users have slightly higher data expenses, 0.49 USD per month and 2.17 USD per month, respectively. The lowest decile for both network users have very low data expenses insufficient for any meaningful use of the internet.
Sixty-eight per cent of total mobile phones were non-smartphone in one survey (Pandey and Raj, 2016). A different survey has similar estimate of non-smartphone users at 62 per cent (Internews, 2014). A third survey in 4,145 households across Nepal puts this number at slightly more than 50 per cent. 10 It is clear that the device people own excludes them from the services and limits their usage even if they were affordable. These surveys put the number of respondents who have used the internet in the last month to within 35 and 50 per cent. There is a caveat. The national regulator of telecommunication in Nepal considers a subscriber with an internet enabled sim-card in their phones as an internet user. This simplistic definition of who qualifies as an internet user is the reason why subscription numbers show a different picture of access than what is revealed by data consumption. Access therefore has to account for both reach and usage.
The ubiquity of mobile phones and the popularity of mobile internet offer a nuanced understanding of inequality in the networked society. Although less surprising, it is also evident in other types of ICT. The number of broadband connection has plateaued with the success of mobile internet. Broadband connection is expensive for the information needs of an average internet user. Again, we could provide various arguments to support this assertion. As Regmi (2017) argues from economic perspective, the household expenditure patterns show it is difficult for the poorest 40 per cent of the households to cut-off their expenses and allocate the surplus to pay the broadband tariff. Without a significant increase in the income levels, it is likely that broadband will remain unaffordable for the majority.
Discussion
Affordability and Inequality
Nepali internet and ICT policies since the 2000s describe access in its simplistic notion of reach. Subscription numbers are used as an alias for adoption. These are evident from the questions asked in large-scale national surveys and across policy narratives. Meanwhile, the understanding of digital divide has moved much beyond the demarcation of people having and not having access to ICT. Mere possession of skills and use are inadequate to determine the beneficial outcomes in the lives of the ICT users, often called the third-level digital divide. Studies have also shown that there are social implications of what people do online (van Deursen & Helsper, 2015).
Even reach does not have a simple policy prescription. One of the reasons behind low adoption numbers of household broadband and limited use of the mobile internet data is affordability. The Broadband Commission (2016) considered a fixed broadband connection as affordable if its tariff does not exceed 5 per cent of monthly per capita income. For mobile broadband, Alliance for Affordable Internet (2018) defines a connection affordable if 1 GB of data costs no more than 2 per cent of monthly per capita income. The latest household budget survey estimates household income of the lowest quintile at NPR 18,338 (USD 177) and highest quintile at NPR 53,578 (USD 518) (Nepal Rastra Bank, 2015). The entry level broadband internet (approx. USD 14 per month) is equivalent to 40 per cent and 13 per cent of the per capita income for the lowest and highest quintile, respectively.
Affordability has a direct relationship with quality-of-service and available bandwidth. The challenge is usually simplified as a direct trade-off between speed and cost. Telecom regulatory authority of India (TRAI) revised the minimum broadband speed from 256 Kbps to 512 Kbps in 2012 (Prasad et al., 2016, pp. 1–6). There was a wide criticism of TRAI buckling under pressure from the telecom operators, who are also internet service providers (ISPs), for a U-turn from the original 2 Mbps (mega-bits per second) recommendation. 11 ISPs stated that without throttling they cannot provide fast internet at an affordable price. Two options were suggested to the regulators, either the minimum was set at 64 kbps or allow unrestricted throttling after quota was exhausted. This Indian experience has shown that owning state of the art infrastructure in a highly competitive market does not necessarily translate to affordable internet even at a modest speed. The minimum broadband speed is also set at 512 kbps in Nepal. This target is neither informed by present or future use of the internet nor with affordability in mind.
A substantial body of literature on digital divide deals with the social consequences of inequality in ICT ownership, access and skills. They are further restricted to age, gender and race distinctions. These dimensions alone cannot capture the complexity in full when we factor in demographic shifts such as rapid rural-to-urban and overseas migration and rising rates of unemployment. Such shifts result in technosocial transformations giving rise to a group that lies between the traditional haves and have-nots, the have-less (Qiu, 2009). The emergence of information have-less is a characteristic of information stratification in poor countries like Nepal (Chautari, 2018). Our estimation in Table 5 showed the top 10 per cent of both NTC and Ncell mobile internet users consumed more than four times data than the remaining 90 per cent during 2016. Only after accounting for the have-less we can explain the curious case of only a small proportion of users consuming large amount of data while there is exponential surge in data traffic. The sheer number of the barely-users account for the large volume of traffic.
Lifting people from have-nots is comparatively easier, as the success of mobile telecommunications has demonstrated. However, majority of them are now in the have-less. Increasingly people in this group with some disposable income to spend on ICTs would rely on their phones for communication and to learn about health care, education, employment and other aspects related to their livelihood. These people are largely ignored by the market and the policymakers. The policies are either aiming to expand connectivity or improve the quality of high-end services. Neither of these is sympathetic to the information needs of the have-less. For instance, when there is a government crackdown on the grey market this directly affects the have-less who are the core consumers for these ICTs. The hyperbole in internet user numbers reported by the telecoms and consequently the relevant authorities comes from measurement problem. They equate ownership of internet enabled sim-cards to internet user. This is precisely why internet policies need to characterise the ICTs of the information have-less and their information needs.
Electricity and Digital Nepal Ambitions
Underdeveloped adjacent infrastructure could have two alarming consequences on the digital infrastructure expansion. First, not all the geographic regions would get the potential benefits of ICT. This side-effect that is even observable in high-income countries in the United Kingdom where e-commerce vendors give less priority to certain areas because of higher delivery costs (Clarke et al., 2015). Second, the future internet and ICT infrastructure would have limited utility in absence of supportive infrastructures like electricity and roads.
Poor access to electricity partially explains the unequal access to digital before and after boom in mobile telecommunications. In the early period, electricity access was one of the important indices that described regions with low and high household ownership of ICT equipment (Pandey & Regmi, 2018). Electricity tariffs could also describe internet access since households with internet access are paying large electricity bills (Pandey & Regmi, 2018). The higher bills do not necessarily result from ICTs consumption but rather explains that internet and ICTs are found in affluent households. Affordability and use as important dimensions of access are clearly visible. Even after 2015, when the mobile subscriptions skyrocketed, access to electricity remained closely tied with access to ICTs. This is also evident from the study by Armey and Hosman (2016). They show a strong relation between electricity and ICT access in the low-income countries. Nepal had a low electricity uptake in the early period with a distinct rural–urban gap. Nepal continued to have a low annual electricity consumption per capita of 139 kWh in 2014, compared to South Asian average of 707 kWh and world average of 3127 kWh. 12 The electricity demand of the envisaged networked economy, ‘Digital Nepal’, is absent from the dominant discourses on connectivity.
Internet usage is energy expensive. Transmitters, amplifiers, routers, web servers, data centres and base transceiver stations demand massive energy. There are millions of end user devices as well. To get an indication of how enormous the power demand can be, a single company Telecom Italia consumed 2 TWh energy in 2006. 13 Adjusting for population difference, this amount of energy is almost half of Nepal’s total electricity consumption in 2017. 14 In addition, the energy demand of the digital infrastructure is increasing globally. An estimation done in 2014 indicated the annual growth in energy consumption by the ICT sector was 10 per cent worldwide, whereas the growth in overall electricity consumption was only 3 per cent (van Heddeghem et al., 2014). There are also some studies that suggest ICT’s energy consumption would saturate instead of increasing continuously (Malmodin et al., 2018). Such estimates of electricity consumption and transmission efficiency vary greatly in the literature. Nonetheless, there is a general agreement that the electricity footprint of the ICT sector is huge, and it will pose a constant challenge for low-income countries where critical issues extend well beyond infrastructure.
In order to understand the implication of ‘Digital Nepal’ policies for the electricity sector, we have to look at the visions and targets set in these documents. Broadband policy 2015 targets to achieve household internet penetration of 40 per cent. ICT policy 2015 aims to provide universal access to broadband by 2020. The grand visions of ICT in these policy documents envisage that ICT and internet will increase productivity, contribute to national GDP, bridge the gap between rural and the urban and reduce inequality by bringing the marginalised groups within the confines of development. These visions touch every aspect of Nepali economy and demand massive infrastructure development efforts to create a ‘knowledge-based economy’. Presently industries in Nepal consume only 8 per cent of the total electricity. 15 The remaining electricity is mostly used for lighting in household. These raise a few straightforward questions. How much electricity is needed to power the digital infrastructure and deliver services from all sectors of the economy? How much electricity would be consumed by the envisaged ICT industry to deliver the growth targets? What is the estimated increase in household electricity consumption when the usage extends beyond lighting? There are no official measurements or even projections. An estimate shows the energy demand of Nepali ICT during 1999–2011 was about one-third of the transportation sector (Regmi & Pandey, 2017).
Electricity deficiency is persisting in Nepal. Nepal had a peak demand of 1,508 MW in the fiscal year 2017/18, whereas the installed capacity was only 1,074 MW (Nepal Electricity Authority, 2018). As the majority of the electricity comes from hydropower projects that are run-of-the-river type, production almost halves in dry seasons (Subedi, 2017). The power deficiency is being managed by cutting electricity to industries, random power cuts and planned power cuts (load shedding). In addition, Nepal has been managing the demand–supply deficit through electricity imports from India. The maximum power import from India in peak hours reached 521 MW in April/May 2018,, which comes mainly from power plants located in the Indian state of Bihar (Nepal Electricity Authority, 2018). This import has led to a temporary relief from load shedding. The temporary nature of the arrangement is also evident as Bihar is deficient in electrical energy (Abhisek et al., 2015). There is optimism that completion of ongoing large hydroelectric projects would be sufficient to sustain electricity during wintertime. We have to remember the development of such hydropower projects may be sufficient only in the business as usual scenario. The energy profile of the present Nepali ICT sector is not known and the energy demand of the ‘Digital Nepal’ is left out as something that can be fixed.
Conclusions
Nepali internet and ICT policies need a nuanced understanding of access that extends beyond reach to technologies. This is not helped by using subscription numbers as an alias for access in these documents. There has to be conceptual shift from the haves and have-nots dichotomy. Through the analysis of data consumption, we argue that majority of the mobile internet users are the have-less. There is a vast middle ground between extending connectivity and improving high-end services. The policies have to address the stratification and the information needs of the have-less. The migrant workers, unemployed youth from lower social strata, older population with limited access to some form of ICTs and the internet have distinctive access, usage and use patterns.
The ‘self-evident’ vision of connectivity offers ‘a powerful, aspatial and ahistorical teleology’ (Friederici et al., 2016). Nepali internet and ICT policies have overlooked the underlying socioeconomic inequalities, the geographical barriers, and the constraints of insufficient supporting infrastructures. This is evident from the trends in imports of ICT goods, which have chased optimism without relevant provisions to capitalise on the new infrastructure. The policies are informed by select discourses and global averages. Very little has been done to understand the ground realities. Importantly, the dominant vision that guides the policies ignores the fact that the process of diffusion of ICTs is carried out within the structural constraints. The haves, have-nots and many gradations in between come into existence from social transformations, which assign distinctive role to communication technologies in their lives and livelihood.
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
This research was done as a part of a two-years project funded by the Ford Foundation. The project was conducted by Martin Chautari, Kathmandu.