## THE EFFECT OF REMITTANCES ON HOUSEHOLD'S EXPENDITURE PATTERNS AND LABOR SUPPLY

## Source details

**Canonical URL:** [THE EFFECT OF REMITTANCES ON HOUSEHOLD'S EXPENDITURE PATTERNS AND LABOR SUPPLY](https://www.imf.org/-/media/files/publications/cr/2020/english/1nplea2020002.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2020/english/1nplea2020002.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2020/english/1nplea2020002.pdf.json)

---

### Introduction
- Remittances play a core role in the Nepalese economy:
  - Workers' remittances reached US$7.8 billion in FY2018/19.
  - This corresponds to a quarter of total GDP, 3 times total exports of goods and services, and about ten times official foreign aid.
  - The annual growth rate of workers' remittances has decreased to an average of 7.5 percent over the past three years, from 23 percent in the 2000s.
  - The number of official monthly new emigrant workers fell to 18 thousand in October 2019, compared to a peak of 62 thousand in July 2014.
  - The outlook for remittances is uncertain in light of the completion of major infrastructure projects, such as those in preparations for the 2020 Dubai World Expo and the 2022 FIFA World Cup in Qatar.
- Study design and data:
  - Uses Nepal Household Risk and Vulnerability Survey – 2016 (NRVS-2016).
  - Employs a propensity score matching (PSM) method to address selection bias and reverse causality.
  - Focuses on effects of remittances on household expenditure patterns and labor supply.

### Stylized Facts: Migrant Workers
- Demographics and migration patterns:
  - Close to 10 percent of total individuals in the HH roster are recorded to be away from home; 21 percent are working in other domestic cities and 79 percent abroad.
  - Destination shares for foreign migrants: GCC 42 percent, India 30 percent, Malaysia 22 percent, other countries (South Korea and Japan) 6 percent.
  - Almost all migrants are male (95 percent).
  - Two-thirds are from rural areas; average age 29.6 years; average stay 23 months.
- Income, costs, and remitting behavior:
  - Average monthly income of migrants (respondents) is NPR 31,376.
  - Average monthly per capita GDP in 2016 noted as NPR 82 thousand (in context: migrants' income is 4.6 times monthly per capita GDP).
  - Average monthly income of foreign migrants: NPR 34,871 (US$328).
  - Average monthly income of domestic migrants: NPR 17,653.
  - Almost all foreign migrants paid migration costs averaging NPR 107,342 (US$1,009), corresponding to three months of their foreign income.
  - 84 percent of foreign migrants remit money to family/relatives/friends.
  - Table 1 observations: total migrants (2,742), domestic migrants (592), foreign migrants (2,150).

### Stylized Facts: Remittance-Receiving Households
- Prevalence and trends:
  - 29 percent of HHs in NRVS-2016 receive remittances (locally or internationally), down from 52 percent in NLSS-2010/11.
  - Domestic remittance-receiving HHs fell to 4 percent in 2016 from 18 percent in 2010; foreign remittance-receiving HHs rose by one percentage point to 25 percent.
- Household characteristics:
  - Heads of remittance-receiving HHs tend to be female and have a lower level of education; more likely to work in agriculture or have no job.
- Income impacts:
  - Average annual income of remittance-receiving HHs before remittances was only 60 percent of that of non-recipient HHs in 2016.
  - With remittances, total income increases to 1.5 times that of non-recipients.
  - Table 3 sample figures (NRVS-2016):
    - Total annual income (All HHs): 123,785
    - Remittances (1 year) (Remittance recipients): 222,390
    - Income + remittances (Remittance recipients): 346,175

### Comparison of Remittance Value (2010/11 vs 2016)
- Per capita HH remittances and per capita nominal GDP ratios (NRVS-2016):
  - Remittance recipients (A) 53,047; Per capita nominal GDP (B) 82,644; Ratio (A/B) 0.642.
  - Foreign remittance recipients (A) 55,606; Ratio (A/B) 0.673.
  - Domestic remittance recipients (A) 34,075; Ratio (A/B) 0.412.
- NLSS-2010/11 ratios:
  - Remittance recipients Ratio (A/B) 0.534; Foreign 0.603; Domestic 0.341.
- Growth of ratio between surveys:
  - 20.1 percent overall; 11.6 percent for foreign; 20.9 percent for domestic.
- Interpretation:
  - Foreign remittances increased from 60 percent of per capita GDP in 2010 to 67 percent in 2016.
  - Domestic remittances increased from 34 percent of per capita GDP in 2010 to 41 percent in 2016.
  - Increase in foreign remittance values linked to higher share of migration to GCC, Japan, and Korea with higher wage levels.

### Effects of Remittances on Households’ Economic Activities — Methodology
- Methodological challenges:
  - Selection bias: remittance-receiving HHs differ systematically (e.g., higher share of female HH heads).
  - Reverse causality: poverty levels may influence remittances as well as be affected by them.
- PSM approach implemented:
  1. Construct a logit model with remittance-receiving HH as binary dependent variable using covariates (share of children/adults/elders, household size, rural area, HH head age, male HH head, HH head education).
  2. Estimate propensity score (probability of getting remittances) for each HH.
  3. Match each remittance-receiving HH with a non-receiving HH of similar propensity score to form artificial control group.
  4. Perform balancing test to check covariate distributions post-matching.
  5. Estimate treatment effects on HH expenditure patterns and labor supply by contrasting treated and control groups.

### Effects on Household Expenditure Patterns
- Key results (foreign remittance focus; NRVS-2016 comparisons):
  - Food consumption share:
    - Foreign remittance recipients: 56.7 percent vs Remittance non-recipients: 61.0 percent — negative and highly significant effect.
  - Non-food daily consumption share:
    - Foreign remittance recipients: 25.6 percent vs non-recipients: 23.7 percent — positive and statistically significant.
  - Durable goods consumption:
    - Foreign remittance recipients: 2.7 percent vs non-recipients: 2.2 percent — positive and statistically significant.
  - Education-related consumption:
    - Foreign remittance recipients: 6.3 percent vs non-recipients: 5.3 percent — positive and statistically significant.
  - Health-related consumption:
    - Foreign remittance recipients: 3.8 percent vs non-recipients: 3.7 percent — small positive effect.
  - Utilities, rent:
    - Foreign remittance recipients: 4.9 percent vs non-recipients: 4.1 percent — positive and statistically significant.
- Aggregate observations:
  - Remittances support greater consumption of productive goods (durable goods, education, health).
  - Total share of productive consumption (durable goods, education, health) remained under 15 percent of total spending.
  - Findings consistent with households treating remittances as temporary income: higher marginal propensity to invest/save out of temporary income leads to higher shares in productive goods but limited change in the overall consumer vs productive goods split absent changes in permanent income.

### Effects on Household Labor Supply
- Analysis:
  - Labor market participation of left-behind HH heads examined using PSM; comparisons done separately for male and female HH heads.
- Results (labor supply comparison, 2016, Table 6):
  - Male head, Foreign remittance recipients:
    - Agriculture 68.0%
    - Non-agriculture 12.4%
    - No job 19.4%
  - Male head, Remittance non-recipients:
    - Agriculture 67.8%
    - Non-agriculture 13.2%
    - No job 18.2%
  - Female head, Foreign remittance recipients:
    - Agriculture 66.7%
    - Non-agriculture 17.7%
    - No job 15.5%
  - Female head, Remittance non-recipients:
    - Agriculture 66.0%
    - Non-agriculture 18.5%
    - No job 14.8%
- Interpretation:
  - No significant difference in unemployment (no job) of household heads between remittance recipients and non-recipients after matching.
  - Statistically significant differences observed prior to matching (20 percent unemployment for remittance recipients vs 13 percent for non-recipients) disappear after PSM.
  - PSM removes confounding bias present in earlier studies that found remittances reduce labor market participation.
  - Results suggest households view remittances as temporary income; HH heads may retain jobs to avoid re-employment difficulties when migrants return.

### Conclusions on Remittances
- Remittances have materially increased household incomes and supported consumption of productive goods (durable goods, education, health).
- Despite higher spending on productive items, total productive consumption remains small (<15 percent of total spending).
- Remittances did not discourage labor supply of left-behind household heads once selection effects are accounted for using PSM.
- Nepal remains vulnerable to declines in remittance inflows given dependence on remittances as a major component of the current account and foreign reserves; recent slowdown in remittance growth and reductions in new emigrant workers underscore uncertainty.

### Remittances — policy options and risks
- Options explored by several studies to reap more benefits from remittances:
  - Lowering foreign migrating costs.
  - Creating incentives so that more remittances flow through formal channels.
- To reduce vulnerability to sudden declines in remittances and the return of foreign migrants, measures needed:
  - Strengthen domestic quality employment opportunities.
  - Incentivize greater private investment by:
    - Removing constraints to FDI.
    - More and better infrastructure investment.
    - Greater competition in product and services markets.
    - Reducing red-tape.

### Infrastructure development and growth — overview
- Potential growth gains:
  - If Nepal were to achieve the average coverage and quality of infrastructure among lower-middle income economies in Asia, its annual growth rate could rise more than 4 percentage points.
  - Such an improvement would require a large investment effort over a number of years.
  - Improving infrastructure would have growth-enhancing effects not fully captured in the exercise; for example, better infrastructure would enable faster industrialization.
- Infrastructure constraints highlighted:
  - Electricity shortages and an inadequate transportation network identified as main constraints to potential growth.
  - Electricity consumption is a twentieth of the global average.
  - Load shedding cost the economy about 7 percent of GDP per year during 2008-2016 until Nepal achieved 24-hour electricity supply in 2018.
  - Road network does not adequately connect tourism areas and the only international airport has exceeded capacity.

### Infrastructure in Nepal compared to regional peers — quantity and quality
- Utilities:
  - Electricity generating capacity increased from 261 Megawatts to about 1100 Megawatt between 1990 and 2017.
  - Current generating capacity per worker remains low, at less than 10 percent of the average capacity among middle-income economies in Asia.
  - Efficiency improved with significant reduction in electricity transmission and distribution losses.
- Transportation:
  - Road network expanded from about 7,000 kilometers in early 1990s to about 30,000 in 2017 (paved and unpaved in total).
  - Road density remains below the average among middle-income economies in Asia.
  - Less than two thirds of roads are paved.
- Telecommunications:
  - Share of households with internet access reached 34 percent in 2017 from 0.2 percent in 2000.
- Aggregate indices:
  - Quantity index: first principal component of electricity generating capacity (Megawatts per 1000 workers), length of roads (kilometer per square meter of land area), and internet access (percent of households with access). First PC weights: electricity 0.61, transport 0.54, telecommunication 0.58; captures 67 percent of overall variation with correlations ranging 0.75−0.83.
  - Quality index: first PC of share of electricity transmitted and distributed (one minus losses), share of paved roads to total, and internet bandwidth per user rescaled 0–1. First PC weights: 0.59, 0.61, and 0.53; captures 51 percent of overall variation with correlations ranging 0.7−0.8.

### Infrastructure and growth — empirical findings
- Empirical approach:
  - Estimate aggregate production function augmented with infrastructure variables following Calderon and Servén (2004, 2008).
  - Controls: human capital (secondary enrollment in percent of total with the age above 15), financial development (domestic credit to private sector in percent of GDP), trade openness (trade in percent of GDP), inflation, government burden (government final consumption expenditure in percent of GDP), institutional quality (ICRG Political Risk Index), terms of trade and their changes, and size of the modern (non-agricultural) sector; all expressed in logs.
  - Panel: 80 countries for 1990−2017 using non-overlapping 5-year period averages.
  - Estimation methods include GMM (Arellano and Bond, 1991) to address endogeneity concerns.
- Key results:
  - Infrastructure indices, both quantity and quality, have a positive and significant relationship with growth.
  - In Nepal, infrastructure expansion—quantity more so than quality—has significantly contributed to higher growth since the mid-1990s.
  - Explosive penetration of telecommunications explains a significant share of infrastructure’s contribution to growth, followed by road expansion and increased electricity generating capacity.
  - Improvement in efficiency of electricity provision has made a positive contribution in terms of infrastructure quality.
  - Beyond infrastructure, lower-middle income economies achieved significant growth through industrialization (increase in share of manufacturing and service sector), which has not yet materialized in Nepal.
- Selected coefficient highlights (Dependent variable: GDP per worker (log difference)):
  - Infrastructure Quantity:
    - 0.034*** (column 1)
    - 0.068*** (column 2)
    - 0.151*** (column 3)
    - 0.216*** (column 4)
    - 0.023** (column 5)
    - 0.042*** (column 6)
    - 0.141*** (column 7)
    - 0.155** (column 8)
  - Infrastructrue [sic] Quality:
    - 0.023*** (column 1)
    - 0.020** (column 2)
    - 0.031** (column 3)
    - 0.055** (column 4)
  - Modern sector share examples:
    - 0.611*** (Within estimator, column 3)
    - Other reported estimates: 0.389***, 0.457***, 0.696***, 1.010** (across specifications)
- Sample and diagnostics:
  - N. observations: 367 (columns 1-3), 270 (column 4), 249 (columns 5-7), 161 (column 8).
  - N. countries: 97, 97, 92, 88 (columns 1-4) and 87, 78, 78, 78 (columns 5-8).
  - N. Instruments: 54 (column 4) and 47 (column 8).
  - Arellano-Bond test for AR(2): 0.10 (column 4), 0.27 (column 8).
  - Hansen test: 0.24 (column 4), 0.42 (column 8).
  - Note: Robust standard errors reported in parentheses. Significance: *** p<0.01, ** p<0.05, * p<0.1. Period dummies & constants are not reported.

### Infrastructure impact on growth and labor productivity
- If Nepal were to achieve the average quantity and quality of infrastructure among lower-middle income economies in Asia, its annual growth rate could rise more than 4 percentage points.
- Sectoral priorities identified:
  - Greatest gaps relative to peers are in utilities sector (electricity generating capacity per worker) and transport sector (expansion of road network and improving share of paved roads).
- Trends and measures:
  - Utility quantity measured as Megawatt per 1000 workers; utility quality measured as electricity transmitted & distributed in percent of total production (Nepal data based on Nepal Electricity Authority estimate for FY 2018/19).
  - Transport quantity measured as road length in Kilometers per square meter; transport quality measured as share of paved roads in percent total road length.
  - Telecommunication quantity measured as percent of households with internet access; telecommunication quality measured as international internet bandwidth per internet user (bit/s).

### Policy implications and public investment management
- Scaling up quality infrastructure development would provide a significant boost to growth and help Nepal achieve its growth ambitions, but would require a large ramp up of infrastructure investment.
- Growth benefits depend crucially on:
  - How additional infrastructure spending is financed: tax increases, government borrowing, or public-private partnerships.
  - How additional infrastructure spending is managed: improving public investment management to raise public investment efficiency and reduce fiscal risks.
- Evidence cited: Countries with less efficient public investment tend to get a lower growth boost from infrastructure spending (IMF, 2015).

### Fintech in Nepal — summary of findings and recommendations
- Fintech advantages and current status (as of January 2020 observations):
  - 18 fintech payment companies licensed by Nepal Rastra Bank: 10 payment system providers and 8 payment system operators.
  - Three international companies including Visa and Mastercard are licensed to operate as payment service providers in Nepal.
  - More than 9 million Nepalese use fintech payment for making transactions; utility payments are the main use of mobile payments.
- Enabling conditions and infrastructure:
  - Electricity supply is increasing and becoming more reliable as hydropower projects come to fruition.
  - Cellular network is expanding to rural areas; number of Nepalese owning cellular lines exceeding 100 percent.
  - Slow speed of cellular data does not appear to be a hindrance for mobile money operation.
- Policy recommendations:
  - Use the Bali Fintech Agenda as a framework to guide policy to promote fintech while mitigating risks.
  - Regulations should support responsible fintech development, address financial integrity risks, facilitate interoperability across service providers and banks, strengthen consumer protection, and expand financial literacy.
  - Regulations should align with the Financial Action Task Force (FATF) standard and may include provisions to allow e-filing and digital verification as part of customer due diligence measures.
  - Address data gaps to better assess fintech expansion and usage; gather data on active usage of formal financial services and identify dormant accounts.
  - Develop a national digital ID to enable expansion from payment services to advanced services such as credit and insurance; reduce paperwork, enable digital data use via APIs, reduce cost and time of AML/CFT customer due diligence, and allow interoperability across digital financial service providers.

*Source: Nepal Household Risk and Vulnerability Survey – 2016, Thapa and Acharya (2017), and IMF staff calculations.*

### References ____________________________________________________________________________________ 11

### THE EFFECT OF REMITTANCES ON HOUSEHOLD'S EXPENDITURE PATTERNS AND LABOR SUPPLY

### Introduction
- Remittances play a core role in the Nepalese economy.
  - Workers' remittances reached US$7.8 billion in FY2018/19.
  - This corresponds to a quarter of total GDP, 3 times total exports of goods and services, and about ten times official foreign aid.
  - The annual growth rate of workers' remittances has decreased to an average of 7.5 percent over the past three years, from 23 percent in the 2000s.
  - The number of official monthly new emigrant workers fell to 18 thousand in October 2019, compared to a peak of 62 thousand in July 2014.
  - The outlook for remittances is uncertain in light of the completion of major infrastructure projects, such as those in preparations for the 2020 Dubai World Expo and the 2022 FIFA World Cup in Qatar.
- Study design and data:
  - Uses Nepal Household Risk and Vulnerability Survey – 2016 (NRVS-2016).
  - Employs a propensity score matching (PSM) method to address selection bias and reverse causality.
  - Focuses on effects of remittances on household expenditure patterns and labor supply.

### Stylized Facts: Migrant Workers
- Demographics and migration patterns:
  - Close to 10 percent of total individuals in the HH roster are recorded to be away from home; 21 percent are working in other domestic cities and 79 percent abroad.
  - Destination shares for foreign migrants: GCC 42 percent, India 30 percent, Malaysia 22 percent, other countries (South Korea and Japan) 6 percent.
  - Almost all migrants are male (95 percent).
  - Two-thirds are from rural areas; average age 29.6 years; average stay 23 months.
- Income, costs, and remitting behavior:
  - Average monthly income of migrants (respondents) is NPR 31,376.
  - Average monthly per capita GDP in 2016 noted as NPR 82 thousand (in context: migrants' income is 4.6 times monthly per capita GDP).
  - Average monthly income of foreign migrants: NPR 34,871 (US$328).
  - Average monthly income of domestic migrants: NPR 17,653.
  - Almost all foreign migrants paid migration costs averaging NPR 107,342 (US$1,009), corresponding to three months of their foreign income.
  - 84 percent of foreign migrants remit money to family/relatives/friends.
  - Table 1 observations: total migrants (2,742), domestic migrants (592), foreign migrants (2,150).

### Stylized Facts: Remittance-Receiving Households
- Prevalence and trends:
  - 29 percent of HHs in NRVS-2016 receive remittances (locally or internationally), down from 52 percent in NLSS-2010/11.
  - Domestic remittance-receiving HHs fell to 4 percent in 2016 from 18 percent in 2010; foreign remittance-receiving HHs rose by one percentage point to 25 percent.
- Household characteristics:
  - Heads of remittance-receiving HHs tend to be female and have a lower level of education; more likely to work in agriculture or have no job.
- Income impacts:
  - Average annual income of remittance-receiving HHs before remittances was only 60 percent of that of non-recipient HHs in 2016.
  - With remittances, total income increases to 1.5 times that of non-recipients.
  - Table 3 sample figures (NRVS-2016):
    - Total annual income (All HHs): 123,785 (implied context of NPRs)
    - Remittances (1 year): 222,390 (Remittance recipients)
    - Income + remittances (Remittance recipients): 346,175

### Comparison of Remittance Value (2010/11 vs 2016)
- Per capita HH remittances and per capita nominal GDP ratios:
  - NRVS-2016 per capita HH remittances (A) and per capita nominal GDP (B):
    - Remittance recipients (A) 53,047; Per capita nominal GDP (B) 82,644; Ratio (A/B) 0.642.
    - Foreign remittance recipients (A) 55,606; Ratio (A/B) 0.673.
    - Domestic remittance recipients (A) 34,075; Ratio (A/B) 0.412.
  - NLSS-2010/11 ratios reported:
    - Remittance recipients Ratio (A/B) 0.534; Foreign 0.603; Domestic 0.341.
  - Growth of ratio between surveys: 20.1 percent overall; 11.6 percent for foreign; 20.9 percent for domestic.
- Interpretation:
  - Foreign remittances increased from 60 percent of per capita GDP in 2010 to 67 percent in 2016.
  - Domestic remittances increased from 34 percent of per capita GDP in 2010 to 41 percent in 2016.
  - Increase in foreign remittance values linked to higher share of migration to GCC, Japan, and Korea with higher wage levels.

### Effects of Remittances on Households’ Economic Activities — Methodology
- Methodological challenges:
  - Selection bias: remittance-receiving HHs differ systematically (e.g., higher share of female HH heads).
  - Reverse causality: poverty levels may influence remittances as well as be affected by them.
- PSM approach implemented:
  1. Construct a logit model with remittance-receiving HH as binary dependent variable using covariates (share of children/adults/elders, household size, rural area, HH head age, male HH head, HH head education).
  2. Estimate propensity score (probability of getting remittances) for each HH.
  3. Match each remittance-receiving HH with a non-receiving HH of similar propensity score to form artificial control group.
  4. Perform balancing test to check covariate distributions post-matching.
  5. Estimate treatment effects on HH expenditure patterns and labor supply by contrasting treated and control groups.

### Effects on Household Expenditure Patterns
- Key results (foreign remittance focus):
  - Food consumption share:
    - Foreign remittance recipients: 56.7 percent (NRVS-2016) vs Remittance non-recipients: 61.0 percent — negative and highly significant effect.
  - Non-food daily consumption share:
    - Foreign remittance recipients: 25.6 percent vs non-recipients: 23.7 percent — positive and statistically significant.
  - Durable goods consumption:
    - Foreign remittance recipients: 2.7 percent vs non-recipients: 2.2 percent — positive and statistically significant.
  - Education-related consumption:
    - Foreign remittance recipients: 6.3 percent vs non-recipients: 5.3 percent — positive and statistically significant.
  - Health-related consumption:
    - Foreign remittance recipients: 3.8 percent vs non-recipients: 3.7 percent — small positive effect.
  - Utilities, rent:
    - Foreign remittance recipients: 4.9 percent vs non-recipients: 4.1 percent — positive and statistically significant.
- Aggregate observations:
  - Remittances support greater consumption of productive goods (durable goods, education, health).
  - Total share of productive consumption (durable goods, education, health) remained under 15 percent of total spending.
  - Findings consistent with households treating remittances as temporary income: higher marginal propensity to invest/save out of temporary income leads to higher shares in productive goods but limited change in the overall consumer vs productive goods split absent changes in permanent income.

### Effects on Household Labor Supply
- Analysis:
  - Labor market participation of left-behind HH heads examined using PSM; comparisons done separately for male and female HH heads.
- Results:
  - No significant difference in unemployment (no job) of household heads between remittance recipients and non-recipients after matching.
  - Statistically significant differences observed prior to matching (20 percent unemployment for remittance recipients vs 13 percent for non-recipients) disappear after PSM.
- Interpretation:
  - PSM removes confounding bias present in earlier studies that found remittances reduce labor market participation.
  - Results suggest households view remittances as temporary income; HH heads may retain jobs to avoid re-employment difficulties when migrants return.

### Conclusions
- Remittances have materially increased household incomes and supported consumption of productive goods (durable goods, education, health).
- Despite higher spending on productive items, total productive consumption remains small (<15 percent of total spending).
- Remittances did not discourage labor supply of left-behind household heads once selection effects are accounted for using PSM.
- Nepal remains vulnerable to declines in remittance inflows given dependence on remittances as a major component of the current account and foreign reserves; recent slowdown in remittance growth and reductions in new emigrant workers underscore uncertainty.

*Source: Nepal Household Risk and Vulnerability Survey – 2016, Thapa and Acharya (2017), and IMF staff calculations.*

### 19.      Remittances have been a key stabilizer of Nepal's weak external sector, but they also

### 19.      Remittances have been a key stabilizer of Nepal's weak external sector, but they also

### Remittances — effects on households and labor supply
- Remittances have improved recipients' standard of living and provided a cushion for economic shocks in Nepal.
- Study: uses the NRVS-2016 survey and a propensity score matching method to analyze effects of remittances on household (HH) expenditure patterns and labor supply.
- Stylized facts from the survey:
  - Migrants are typically young male migrants (30 years old and from urban areas).
  - Typical duration abroad: two years.
  - Heads of remittance-receiving HHs tend to be female with a low education level.
  - Heads of remittance-receiving HHs tend to work in agriculture or have no job.
- Impact on expenditure:
  - Remittance-receiving households spend less on consumer goods (food).
  - Remittance-receiving households spend more on productive goods (durable goods, education, and health).
  - Interpretation: remittances are being treated as temporary income.
- Impact on labor supply:
  - Contrary to some other research, this study shows that remittance-receiving HHs do not decrease their labor supply in Nepal.

### Remittances — policy options and risks
- Options explored by several studies to reap more benefits from remittances:
  - Lowering foreign migrating costs.
  - Creating incentives so that more remittances flow through formal channels.
- To reduce vulnerability to sudden declines in remittances and the return of foreign migrants, measures needed:
  - Strengthen domestic quality employment opportunities.
  - Incentivize greater private investment by:
    - Removing constraints to FDI.
    - More and better infrastructure investment.
    - Greater competition in product and services markets.
    - Reducing red-tape.

### Labor supply comparison (2016) — key figures from Table 6
- Male head, Foreign remittance recipients:
  - Agriculture 68.0%
  - Non-agriculture 12.4%
  - No job 19.4%
- Male head, Remittance non-recipients:
  - Agriculture 67.8%
  - Non-agriculture 13.2%
  - No job 18.2%
- Female head, Foreign remittance recipients:
  - Agriculture 66.7%
  - Non-agriculture 17.7%
  - No job 15.5%
- Female head, Remittance non-recipients:
  - Agriculture 66.0%
  - Non-agriculture 18.5%
  - No job 14.8%
- Source cited for table: Nepal Household Risk and Vulnerability Survey (2016), and IMF staff calculations.

### Infrastructure development and growth — overview
- Scaling up quality infrastructure investment can support Nepal’s growth objectives by expanding access to basic services, jobs, and markets, and boosting private sector productivity.
- If Nepal were to achieve the average coverage and quality of infrastructure among lower-middle income economies in Asia, its annual growth rate could rise more than 4 percentage points.
- Infrastructure constraints highlighted:
  - Electricity shortages and an inadequate transportation network identified as main constraints to potential growth.
  - Electricity consumption is a twentieth of the global average.
  - Load shedding cost the economy about 7 percent of GDP per year during 2008-2016 until Nepal achieved 24-hour electricity supply in 2018.
  - Road network does not adequately connect tourism areas and the only international airport has exceeded capacity.

### Infrastructure in Nepal compared to regional peers — quantity and quality
- Utilities:
  - Electricity generating capacity increased from 261 Megawatts to about 1100 Megawatt between 1990 and 2017.
  - Current generating capacity per worker remains low, at less than 10 percent of the average capacity among middle-income economies in Asia.
  - Efficiency improved with significant reduction in electricity transmission and distribution losses.
- Transportation:
  - Road network expanded from about 7,000 kilometers in early 1990s to about 30,000 in 2017 (paved and unpaved in total).
  - Road density remains below the average among middle-income economies in Asia.
  - Less than two thirds of roads are paved.
- Telecommunications:
  - Share of households with internet access reached 34 percent in 2017 from 0.2 percent in 2000.
- Aggregate indices:
  - Aggregate synthetic indices constructed as the first principal component of electricity generating capacity (Megawatts per 1000 workers), length of roads (kilometer per square meter of land area), and internet access (percent of households with access).
  - First PC weights on electricity, transport, and telecommunication variables: 0.61, 0.54, and 0.58 respectively; captures 67 percent of overall variation with correlations ranging 0.75−0.83.
  - Quality index constructed as first PC of share of electricity transmitted and distributed (one minus losses), share of paved roads to total, and internet bandwidth per user rescaled 0–1; first PC weights 0.59, 0.61, and 0.53 respectively; captures 51 percent of overall variation with correlations ranging 0.7−0.8.

### Infrastructure and growth — empirical findings
- Approach:
  - Estimate an aggregate production function augmented with infrastructure variables following Calderon and Servén (2004, 2008).
  - Controls include human capital (secondary enrollment in percent of total with the age above 15), financial development (domestic credit to private sector in percent of GDP), trade openness (trade in percent of GDP), inflation, government burden (government final consumption expenditure in percent of GDP), institutional quality (ICRG Political Risk Index), terms of trade and their changes, and size of the modern (non-agricultural) sector, all expressed in logs.
  - Panel: 80 countries for 1990−2017 using non-overlapping 5-year period averages.
  - Estimation methods include GMM (Arellano and Bond, 1991) to address endogeneity concerns.
- Results:
  - Infrastructure indices, both quantity and quality, have a positive and significant relationship with growth.
  - In Nepal, infrastructure expansion—quantity more so than quality—has significantly contributed to higher growth since the mid-1990s.
  - The explosive penetration of telecommunications explains a significant share of infrastructure’s contribution to growth, followed by road expansion and increased electricity generating capacity.
  - Improvement in efficiency of electricity provision has made a positive contribution in terms of infrastructure quality.
  - Beyond infrastructure, lower-middle income economies achieved significant growth through industrialization (increase in share of manufacturing and service sector), which has not yet materialized in Nepal.

*Source: Nepal Household Risk and Vulnerability Survey (2016), and IMF staff calculations.*

### 11.      The results suggest that if Nepal were

### 1nplea2020002 - 11.      The results suggest that if Nepal were

### Infrastructure impact on growth and labor productivity
- If Nepal were to achieve the average quantity and quality of infrastructure among lower-middle income economies in Asia, its annual growth rate could rise more than 4 percentage points.
- Such an improvement would require a large investment effort over a number of years.
- Improving infrastructure would have growth-enhancing effects not fully captured in the exercise; for example, better infrastructure would enable faster industrialization, identified as an important contributing factor enabling a growth leap among lower-middle income economies.

### Sectoral priorities identified
- Greatest gaps relative to peers are in:
  - Utilities sector: electricity generating capacity per worker.
  - Transport sector: expansion of road network and improving the quality of road network (share of paved roads).
- Telecommunication quality and quantity, transportation quality and quantity, and utilities quality and quantity each contribute to labor productivity growth benefits (contribution presented in percentage points in staff estimates).

### Empirical findings and robustness
- Infrastructure quantity and infrastructure quality both show significant positive relationships with GDP per worker (log difference) across estimators.
- Selected coefficient estimates from Table 1 (Dependent variable: GDP per worker (log difference)):
  - Lag. Output: -0.061*** (Pooled OLS), -0.106*** (Panel with time effects), -0.377*** (Within estimator), -0.379*** (Difference GMM) — and similar negative significant estimates in columns (5)-(8).
  - Education: 0.043*** (Pooled OLS) and varied estimates across specifications.
  - Modern sector share: 0.611*** (Within estimator, column 3) and 0.389***, 0.457***, 0.696***, 1.010** (columns reported for alternative specifications).
  - Infrastructure Quantity: 0.034*** (column 1), 0.068*** (column 2), 0.151*** (column 3), 0.216*** (column 4); 0.023** (column 5), 0.042*** (column 6), 0.141*** (column 7), 0.155** (column 8).
  - Infrastructrue [sic] Quality: 0.023*** (column 1), 0.020** (column 2), 0.031** (column 3), 0.055** (column 4).
- Sample and diagnostic details:
  - N. observations: 367 (columns 1-3), 270 (column 4), 249 (columns 5-7), 161 (column 8).
  - N. countries: 97, 97, 92, 88 (columns 1-4) and 87, 78, 78, 78 (columns 5-8).
  - N. Instruments: 54 (column 4) and 47 (column 8).
  - Arellano-Bond test for AR(2): 0.10 (column 4), 0.27 (column 8).
  - Hansen test: 0.24 (column 4), 0.42 (column 8).
- Note: Robust standard errors reported in parentheses. Significance: *** p<0.01, ** p<0.05, * p<0.1. Period dummies & constants are not reported.

### Trends in infrastructure (1990-2015 and latest)
- Utility infrastructure — electricity generating capacity measured in Megawatt per 1000 workers (simple average within a group); reference year is the latest available year, 2016-17 in most cases.
- Utility infrastructure quality measured as electricity transmitted & distributed to consumers in percent of total production. Data for Nepal is based on the estimate from the Nepal Electricity Authority for FY 2018/19.
- Transport infrastructure — road length measured in Kilometers per square meter (simple average within a group); transport quality measured as share of paved roads in percent total road length.
- Telecommunication infrastructure — Internet access measured as percent of households with internet access; telecommunication quality measured as international internet bandwidth per internet user (bit/s).
- Group labels: NPL (Nepal), LMI-Asia (lower-middle income Asia), UMI-Asia (upper-middle income Asia). Each bar in figures shows the relevant period average among countries in the group.

### Policy implications and public investment management
- Scaling up quality infrastructure development would provide a significant boost to growth and help Nepal achieve its growth ambitions, but would require a large ramp up of infrastructure investment.
- Growth benefits depend crucially on:
  - How additional infrastructure spending is financed: tax increases, government borrowing, or public-private partnerships.
  - How additional infrastructure spending is managed: improving public investment management to raise public investment efficiency and reduce fiscal risks.
- Evidence cited: Countries with less efficient public investment tend to get a lower growth boost from infrastructure spending (IMF, 2015).

### Fintech in Nepal — summary of findings and recommendations
- Fintech offers advantages for Nepal, such as overcoming geographical barriers and promoting greater mobile payments by tourists.
- Key observations as of January 2020:
  - 18 fintech payment companies licensed by Nepal Rastra Bank: 10 payment system providers and 8 payment system operators.
  - Three international companies including Visa and Mastercard are licensed to operate as payment service providers in Nepal.
  - More than 9 million Nepalese use fintech payment for making transactions; utility payments are the main use of mobile payments.
- Enabling conditions and infrastructure:
  - Electricity supply is increasing and becoming more reliable as hydropower projects come to fruition.
  - Cellular network is expanding to rural areas; number of Nepalese owning cellular lines exceeding 100 percent.
  - Slow speed of cellular data does not appear to be a hindrance for mobile money operation.
- Policy recommendations:
  - Use the Bali Fintech Agenda as a framework to guide policy to promote fintech while mitigating risks.
  - Regulations should support responsible fintech development, address financial integrity risks, facilitate interoperability across service providers and banks, strengthen consumer protection, and expand financial literacy.
  - Regulations should align with the Financial Action Task Force (FATF) standard and may include provisions to allow e-filing and digital verification as part of customer due diligence measures.
  - Address data gaps to better assess fintech expansion and usage; gather data on active usage of formal financial services and identify dormant accounts.
  - Develop a national digital ID to enable expansion from payment services to advanced services such as credit and insurance; reduce paperwork, enable digital data use via APIs, reduce cost and time of AML/CFT customer due diligence, and allow interoperability across digital financial service providers.

*Source: IMF staff estimates and analysis as presented in the provided PDF content unit.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2020/english/1nplea2020002.pdf_