The Digital Divide and Digital Nepal

Two students in the same country sit the same national examination.

One lives in Kathmandu with fibre broadband, a laptop and parents who use online banking; the other lives in a Karnali village where the 4G signal reaches only the hilltop, electricity cuts out for days, and the only phone is shared by the family.

Both are "connected" in the national statistics, yet their digital lives are of different kinds.

The digital divide names this gap, and this unit shows how to define it, how to measure it (shares, ratios, a Lorenz curve with a Gini coefficient, a composite index), and what Nepal's policy response, the Digital Nepal Framework, tries to do about it.

All province-level figures in this unit are invented for teaching so that the arithmetic can be followed; real figures come from the National Statistics Office, the Nepal Telecommunications Authority (NTA) monthly MIS reports and the Ministry of Communication and Information Technology. Always check the latest publication.

3.1 What is the digital divide?

The digital divide is the gap between those who have effective access to digital technologies and those who do not, and between those who benefit from them and those who do not. Researchers describe three levels:

1. First-level divide (access): devices, connectivity, electricity, affordability.
2. Second-level divide (skills and use): digital literacy, language, confidence; the difference between scrolling a feed and filing a tax return.
3. Third-level divide (outcomes): who actually gains income, education, health or political voice from being online.

Closing the first level does not close the others: a smartphone gives access, but a farmer who cannot read the interface in her language or does not trust mobile money gains little.

Dimension	How the gap appears in Nepal
Geography	Urban valleys and the Terai towns have fibre and 4G; remote hill and mountain districts have thin coverage and costly backhaul
Income and cost	Data and device costs are a larger share of a low income; second-hand phones; no home computer
Gender	Women have fewer personal phones and lower use; norms, literacy and safety concerns
Age and disability	Older people and people with disabilities find interfaces and services inaccessible
Language and script	Most content is English; Nepali Unicode support, voice and local-language interfaces are limited; many mother-tongue communities have almost none
Caste and ethnicity	Historical disadvantage compounds access barriers
Skills	Low digital and financial literacy limits safe use

3.2 Measuring the divide

Simple ratios and gaps

Compare two groups with a gap (difference in percentage points) or a ratio (relative disadvantage).

Suppose 38% of women and 52% of men use the internet: the gap is 14 points and the gender parity index (female ÷ male) is 0.731; parity is 1.0.

Reporting both avoids mistakes: a 14-point gap sounds small if both groups are near 90%, but is large when use is 38% versus 52%.

Totals: population 29.07 million, of whom 16.52 million use the internet, a national rate of 56.8%. Bagmati's internet-use rate is 2.93 times Karnali's, and its fixed broadband penetration is 10.0 times higher.

The national average hides these differences, which is why disaggregated data matter.

Lorenz curve and Gini coefficient

Borrowed from the study of income inequality, the Lorenz curve plots the cumulative share of the population (poorest access first) against the cumulative share of internet users.

Perfect equality is the diagonal; the farther the curve bends below it, the greater the inequality.

The Gini coefficient is the area between the diagonal and the curve divided by the area under the diagonal: G = 1 − Σ (Xᵢ − Xᵢ₋₁)(Yᵢ + Yᵢ₋₁), summing over groups with cumulative population share X and cumulative user share Y. 0 is equality; 1 is one group having everything. Sorting provinces by users per head:

Province (poorest access first)	Population share	Internet-user share	Cumulative population	Cumulative users
Karnali	5.8%	2.9%	5.8%	2.9%
Sudurpashchim	9.3%	5.9%	15.1%	8.8%
Madhesh	21.0%	15.1%	36.1%	23.9%
Lumbini	17.5%	17.0%	53.7%	40.9%
Koshi	17.1%	18.7%	70.8%	59.5%
Gandaki	8.3%	10.2%	79.0%	69.7%
Bagmati	21.0%	30.3%	100.0%	100.0%

The Gini is 0.171. This is low because the province is a coarse grouping; the divide inside provinces (urban against rural municipalities) and between households would be larger. Always state the unit of analysis before comparing Gini values.

A composite digital divide index

No single indicator captures the divide, so analysts combine several.

A common method is min-max normalisation: for each indicator, score = (value − minimum) / (maximum − minimum), so the lowest province gets 0 and the highest 1, then average the scores with equal weights (or chosen weights).

Province	Internet	Broadband	Smartphone	Skills	DDI (mean of 4)
Koshi	0.63	0.47	0.52	0.46	0.520
Madhesh	0.24	0.14	0.24	0.10	0.181
Bagmati	1.00	1.00	1.00	1.00	1.000
Gandaki	0.78	0.58	0.74	0.62	0.681
Lumbini	0.50	0.31	0.46	0.38	0.409
Karnali	0.00	0.00	0.00	0.00	0.000
Sudurpashchim	0.15	0.06	0.13	0.15	0.120

Bagmati ranks first (index 1.000) and Karnali last (0.000). Min-max scores are relative: a score of 0 means "lowest of these seven", not "no access". Changing the indicators or weights can change the middle ranks; publish the method with the result.

3.3 Language, script and content

A person cannot benefit from technology that does not speak their language.

Unicode encodes Devanagari so that Nepali text, with its conjuncts and vowel signs, is stored and searched consistently; before Unicode, legacy fonts such as Preeti stored text in ways incompatible with search, email and the web, leaving much older government content unsearchable.

Needs: Nepali keyboard layouts (Romanized and traditional), fonts that render conjuncts correctly, spell-checkers and OCR for Devanagari, text-to-speech and speech recognition in Nepali and other languages, and interfaces and government forms in the languages of the country (Nepal has over a hundred languages).

Local content (agriculture advice, health information, news, education material) increases the relevance of connectivity. Accessibility standards (WCAG: text alternatives, contrast, keyboard operation, captions) address disability.

3.4 The Digital Nepal Framework

The Government of Nepal launched the Digital Nepal Framework in 2019 as a national roadmap to use digital technology for economic growth and better services.

It identifies eight sectors where digital technology can have the largest effect and a set of enabling "digital foundation" building blocks:

Sector	Examples of the digital focus
Digital Foundation	Connectivity and broadband, national data centre and cloud, digital ID, digital payments, cyber security, legal and regulatory reform, skills
Agriculture	Information and advisory services for farmers, market prices, supply-chain digitisation, e-agriculture
Health	Health information systems, electronic records, telemedicine
Education	Digital learning content, connectivity of schools, e-learning platforms
Energy	Smart metering, digital management of generation and distribution, electric mobility planning
Tourism	Online services for visitors, digital promotion, reservation and information systems
Finance	Digital payments and financial inclusion, interoperability, digital banking
Urban Infrastructure	Smart city and municipal digital services, traffic and utility management

The framework lists many initiatives (on the order of eighty) with agencies and timelines; its progress depends on funding, coordination across ministries and the three tiers of government, and the legal reforms discussed in Unit 4.

Complementary programmes: community information centres and community learning centres, school computer-lab and connectivity programmes, rural telecommunication development funded through the universal service fund of the NTA, and the expansion of optical-fibre backbones by the national utilities and private operators.

Evaluate any programme by the three divides: did it provide access, build skills, and produce outcomes?

3.5 Closing the divide

• Infrastructure and affordability: shared towers, universal service funding, tax treatment of devices and data, community Wi-Fi, solar-powered access points where the grid is weak.
• Skills: digital literacy in schools and in community centres, training for women and older people, train-the-trainer networks.
• Content and services: services that work in Nepali and local languages, on low-end phones, offline or by USSD/SMS, with human help desks.
• Inclusive design: consult the excluded, test with real users, accessibility by default.
• Trust and safety: consumer protection for digital money and clear complaint channels; safe online spaces for women and girls.
• Measurement: collect gender-, age-, location- and disability-disaggregated data and publish it.

3.6 Seeing the ideas: the infrastructure, the three levels, the Lorenz curve and the index

The divide starts with physical infrastructure. Plate 3.1 shows what connectivity is made of: a fibre cable that carries the bandwidth, a router that shares it, a basic handset and a smartphone, the devices at the end of the line.

(a) Fibre backbone

(b) Broadband router

(c) Feature phone

(d) Smartphone

3.6.1 Eight ways to close the digital divide

1. Extend networks to remote areas. Fibre, towers and community networks must reach hill and mountain districts. The cost per user is high, so a shared universal service fund helps. Karnali needs connections before apps.
2. Make devices and data affordable. Lower taxes on basic smartphones and cheaper data packs widen access. Instalment plans and refurbished devices help low-income families. Public Wi-Fi in wards and libraries serves those without a plan.
3. Provide reliable electricity. A phone or router is useless without power. Solar charging points in remote schools and health posts keep services running. Power and connectivity must be planned together.
4. Teach digital skills. Schools, campuses and ward offices can offer short practical courses. A volunteer who shows a grandmother how to scan a QR code does real good. Skills lead to safe and confident use.
5. Create Nepali and local-language content. Interfaces, voice assistants and forms in Nepali and mother tongues remove the language barrier. Unicode support must work on cheap phones. People trust what they can read.
6. Include women, older people and persons with disabilities. Design for screen readers, large text and voice. Women-only training groups and safe reporting channels raise confidence. Inclusive design benefits everyone.
7. Offer e-services that save time. Online tax, licence and certificate services give people a reason to join. Service centres with trained staff help those who cannot use them alone. Real benefits drive demand.
8. Measure and publish the gaps. Collect data by province, gender, age and disability, and compute gaps, ratios and indices. Public numbers keep governments accountable. What is measured is more likely to be fixed.

Chapter summary

• The divide has three levels (access, skills, outcomes) and many dimensions: geography, income, gender, age, disability, language and caste.
• Gaps, ratios, the gender parity index, Lorenz/Gini and min-max composite indices measure it; the unit of analysis and the method must be stated.
• Unicode, Nepali-language tools and local content are conditions for benefit, not extras.
• The Digital Nepal Framework (2019) has eight sectors: Digital Foundation, Agriculture, Health, Education, Energy, Tourism, Finance and Urban Infrastructure.
• Closing the divide needs infrastructure, affordability, skills, local content, inclusive design and trust, and assisted channels.

Practice

1. Name the three levels of the digital divide.
2. State two ways the divide appears for women in Nepal.
3. With 45% of women and 60% of men online, compute the gap and the gender parity index.
4. In the example data, how many times higher is Madhesh's internet-use rate than Karnali's? (Use the table.)
5. Explain how the Lorenz curve and Gini coefficient are computed.
6. Why is a min-max index a relative, not absolute, measure?
7. Why did legacy fonts such as Preeti harm digital inclusion?
8. List the eight sectors of the Digital Nepal Framework.
9. Name four measures to reduce the divide in a remote municipality.
10. Why should e-government keep an assisted channel?

Answers and worked solutions

1. Access, skills/use, outcomes.

2. Fewer women own a personal phone; lower literacy and safety concerns restrict use (also family norms and content not in their language).

3. Gap = 15 points; parity index = 45/60 = 0.75.

4. 41 / 28 = 1.46 times.

5. Sort groups by access per head; plot cumulative population share against cumulative user share; Gini = 1 − Σ(Xᵢ − Xᵢ₋₁)(Yᵢ + Yᵢ₋₁), the area between the diagonal and the curve over the area below the diagonal.

6. The scores rank places against the best and worst in the sample; a score of 0 only means lowest of those places.

7. Their encoding was not standard, so text could not be searched, copied between systems or displayed reliably on the web or phones.

8. Digital Foundation, Agriculture, Health, Education, Energy, Tourism, Finance, Urban Infrastructure.

9. Shared or community connectivity (solar-powered), digital-literacy training, services in local language and offline/USSD, subsidised devices or community access centres.

10. People without access or skills would otherwise be excluded from essential services.

Model answers: medium (5 marks) and long (10 marks) questions

Q1 (5 marks). What is the digital divide? Explain its three levels with Nepali examples.

Introduction. The digital divide is the gap between those who have effective access to digital technologies and those who do not, and between those who benefit from them and those who do not (section 3.1 and Figure 3.1).

1. First level: access. Devices, connectivity, electricity and affordability. Example: a remote village in Karnali with thin 4G coverage and costly data.
2. Second level: skills and use. Digital literacy, language, confidence and trust. Example: a farmer who has a smartphone but cannot read the English interface or does not trust mobile money.
3. Third level: outcomes. Who actually gains income, learning, health or political voice. Example: two students with the same phone, only one of whom uses it to learn and earn.

Level	Barrier	Possible remedy
Access	No signal, cost	Networks, subsidies
Skills	Language, fear	Training, Nepali content
Outcomes	No useful services	E-services, jobs

Dimensions. The gap appears along geography, income, gender, age and disability, language and script, caste and ethnicity, and skills.

Conclusion. Closing level 1 does not close levels 2 and 3, so policy must address all three.

Q2 (5 marks). Explain how the digital divide is measured with gap and ratio. Use the example data to compute the national internet-use rate and compare Bagmati with Karnali.

Introduction. Measurement turns a vague concern into numbers that can be tracked (section 3.2).

Gap and ratio. A gap is a difference in percentage points and a ratio shows relative disadvantage. If 38% of women and 52% of men use the internet, the gap is 14 points and the gender parity index is 38 / 52 = 0.731 (parity is 1.0).

Report both, because a 14-point gap means different things at 90% and at 38%.

Provincial data (invented). The seven provinces have 29.07 million people, of whom 16.52 million use the internet, a national rate of 16.52 / 29.07 = 56.8%.

Indicator	Bagmati	Karnali	Ratio
Internet use	82%	28%	2.93
Fixed broadband per 100 homes	40	4	10.0
Smartphone	70%	24%	2.92

Interpretation. The national average of 56.8% hides a provincial range from 28% to 82%, so disaggregated data are essential. The Lorenz curve and Gini coefficient (0.171 for provinces) summarise the inequality, but a coarse unit hides divides inside provinces.

Q3 (10 marks). Explain how a composite digital divide index is built using min-max normalisation. Work out the index for Madhesh from the example data and discuss the limits of such an index.

Introduction. No single indicator captures the digital divide, so analysts combine several into a composite index. Min-max normalisation makes indicators with different units comparable (section 3.2 and Figure 3.3).

Method.

1. Choose indicators: internet use, fixed broadband per 100 homes, smartphone share and digital-skills share.
2. For each indicator, score = (value - minimum) / (maximum - minimum), so the lowest province gets 0 and the highest gets 1.
3. Average the four scores with equal weights, or with chosen weights.
4. Rank the units and publish the method.

Worked example for Madhesh. Minimum and maximum values come from Karnali and Bagmati.

Indicator	Value	Min	Max	Score
Internet use %	41	28	82	(41 - 28) / 54 = 0.241
Broadband per 100	9	4	40	(9 - 4) / 36 = 0.139
Smartphone %	35	24	70	(35 - 24) / 46 = 0.239
Skills %	25	20	68	(25 - 20) / 48 = 0.104

The index is (0.241 + 0.139 + 0.239 + 0.104) / 4 = 0.723 / 4 = 0.181, which matches the table in section 3.2. Bagmati scores 1.000, Karnali 0.000, Gandaki 0.681, Koshi 0.520, Lumbini 0.409 and Sudurpashchim 0.120, so Madhesh ranks fifth of seven.

Use. The index shows where to invest first, here Karnali, Sudurpashchim and Madhesh, and the four columns show what kind of help each needs.

For example, Madhesh's broadband score of 0.139 is lower than its internet score of 0.241, which points to fixed-line infrastructure.

Limits. Min-max scores are relative: 0 means the lowest of these seven provinces, not no access. Changing the indicators, the weights or the units can change the middle ranks, and province averages hide divides inside provinces.

The data in the book are invented for teaching, so real values from NTA and census sources must be used for policy.

Conclusion. A composite index is a useful summary if the method is transparent and the user remembers that the scores are relative.