The Role of Remittances in Household Spending in Rural Nepal

Abstract

Foreign remittances have become a crucial component of the Nepalese economy. This study investigates the impact of remittances on household consumption patterns in rural Nepal using data from the World Bank’s Nepal Household Risk and Vulnerability (NHRV) Survey Panel, covering the period from 2016 to 2018. Employing an instrumental variable regression approach, we estimate the elasticity of remittances to various consumption categories. Our findings indicate that foreign remittances significantly affect total consumption expenditure. Disaggregated results reveal that remittances positively influence spending on food items and non-food categories such as education and healthcare, highlighting their role in enhancing nutrition and human capital development. However, remittances do not contribute to unproductive expenditures like tobacco, alcohol, or rituals. Therefore, other things remaining the same, remittance is enhancing welfare in rural Nepali households.

1. Introduction

Disparities in economic opportunities and earnings have driven cross-country labor migration (Mishra et al., 2022), a trend further reinforced by advances in digital connectivity, migration networks, and transportation (Ghodsi et al., 2024; Munshi, 2003). The international migration of workers has increased over time, rising from approximately 173 million (2.8% of the global population) in 2000 to 281 million (3.6% of the global population) by 2020. OECD countries alone absorbed about 40% of these migrants, while other lower–middle-income countries, including Gulf nations, also attracted a significant share (IOM, 2022; World Bank, 2023b).

As international migration has flowed from low-income to high-income countries, remittance receipts have flowed in the opposite direction. In 2022 alone, global remittance receipts reached a record high of 831 billion USD, marking a 4.81% annual increase. Over three-quarters of this amount went to low- and middle-income countries. For many countries, the outflux of labor and the influx of remittances are crucial to livelihoods and development (IOM, 2022; World Bank, 2023b, 2024). Remittance receipts serve as the primary source of foreign exchange in developing countries. In the global context, the South Asian region remains a major recipient, receiving close to a quarter of global remittances (World Bank, 2023a).

In this context, Nepal stands out as one of the most notable recipients of remittances. In 2022, remittances alone covered 56.6% of Nepal’s trade deficit and were a crucial component of the country’s balance of payments (NRB/GoN, 2022). In terms of the remittance-to-GDP ratio, Nepal ranks ninth globally (World Bank, 2023a), fifth among low- and middle-income countries, and first in South Asia (World Bank, 2024). While the volume of remittance inflows may be relatively small due to the country’s population size and economy, the share of remittances in Nepal’s GDP remains remarkably high. For decades, remittances have consistently accounted for about a quarter of Nepal’s GDP (NRB/GoN, 2022), making them an indispensable factor in development planning and policy-making in the country.

Research on the impact of remittances at both the household and country levels in South Asia reveals varied findings. For instance, De and Ratha (2012) found that remittances positively affect child education in Sri Lanka but do not significantly influence overall consumption expenditure. On the other hand, Raihan et al. (2022) conducted a disaggregated analysis in Bangladesh, concluding that remittances positively impact expenditures on housing, health, land, education, and investment.

At the household level in Nepal, remittances play a crucial role in financing healthcare, education, and other consumption expenses, becoming an essential component of development strategies. Raut and Tanaka (2018) found that the absence of a father—often the household head—negatively affects a child’s education, but remittances help compensate for this by contributing to capital formation. In contrast, Bansak and Chezum (2009) and Bansak et al. (2015) showed that households receiving remittances saw improved educational outcomes and increased human capital accumulation.

Additionally, Mishra et al. (2022) identified a positive relationship between household expenditure and remittances, particularly for food and education, while finding an inverse relationship with tobacco consumption. Remittances have also benefited left-behind family members, particularly women, improving their well-being (Maharjan et al., 2012), increasing household healthcare usage (Kapri & Jha, 2020), and enhancing the overall budgetary status (Mishra et al., 2022). Furthermore, studies by Shrestha (2017), Lokshin et al. (2010), and Wagle and Devkota (2018) suggest that remittances have contributed to poverty reduction in Nepal. For example, Shrestha (2017) found that migration to Gulf countries and Malaysia alone helped reduce poverty by 40% between 2001 and 2011.

In terms of national economic impact, Loser et al. (2006) examine how remittance receipts, as international transfer payments, interact with a country’s fiscal, monetary, and balance-of-payments variables. As remittances increase, disposable income rises, which can influence the exchange rate mechanism, potentially leading to an inverse effect on the recipient country. This effect can manifest directly through increased demand for ’tradable’ goods such as exports and imports or shifts in relative prices (Fullenkamp et al., 2008). In simpler terms, as remittances boost domestic demand, local prices and wages may rise, leading to an appreciation of the local currency. Consequently, this can put pressure on both external and internal sectors, often in an undesirable direction (Thapa-Parajuli, 2013).

Remittances are vital for development and economic growth in developing countries (Cazachevici et al., 2020; World Bank, 2024), stimulating economic activity by increasing both consumption and investment. However, although remittance income boosts household consumption, it does not necessarily channel to productive investments, especially in the case of Nepal (Thapa-Parajuli, 2013). Additionally, remittances can lead to the Dutch disease effect, where an inflow of foreign currency causes an appreciation of the local currency, potentially harming the country’s export sector. They may also foster policy laxity among decision-makers, discouraging improvements in the investment climate (Sapkota, 2013). There remains much to explore about the scale and quality of international migration, remittance receipts, and their economic implications for developing countries in general, and for Nepal in particular.

Classical political economy, drawing on the seminal works of Smith, Mill, and Marx, critically examines sustainable development through a historical perspective, interdisciplinary approaches, and social class analysis (Manioudis & Meramveliotakis, 2022). While these theories may be regarded as historically rooted, they continue to offer valuable methodological insights for contemporary development research. However, the predominant framework for current development interventions is the Sustainable Development Goals (SDGs), which are united under the broader framework of the Millennium Development Goals (MDGs). These goals address the fundamental challenges humanity must confront, not only to achieve sustainable development but to ensure survival on Earth (Klarin, 2018). Although the SDG framework does not explicitly incorporate international remittances, they appear to intersect with several SDGs, influencing global development outcomes in complex and interconnected ways.

While not explicitly included in the Sustainable Development Goals (SDGs), remittances play a significant role in supporting the achievement of several of these goals. Initially, they were not envisioned as primary drivers for SDG attainment. However, empirical evidence highlights their substantial contributions to poverty reduction (SDG 1) (Malecki, 2021), alleviating hunger (SDG 2) (Subramaniam et al., 2021), promoting health and well-being (SDG 3) (Nasrin et al., 2024), and enhancing education quality (SDG 4) in Nepal (Bansak et al., 2015). Additionally, remittances contribute to gender equality (SDG 5) (Nasrin et al., 2024), foster economic growth (SDG 8) (Yerrabati, 2024), reduce inequality (SDG 10) (Azizi, 2019), support climate response efforts (SDG 13) (Azizi, 2019), and strengthen global partnerships (SDG 17) (Akanle et al., 2022).

In Nepal, there are contrasting arguments regarding the impact of remittances. This debate spans two key dimensions: the livelihood approach versus the macroeconomic approach, and productive versus unproductive expenditure. From a microeconomic perspective, predominantly aligned with the livelihood approach, remittances are praised for significantly improving household access to food, health, education, and physical and human capital formation (Mishra et al., 2022). However, even proponents of this perspective acknowledge the social costs of remittances, such as family separations and related challenges (Raut & Tanaka, 2018). From a macroeconomic point of view, remittances are valued as the main source of foreign exchange (Sapkota, 2013). However, concerns such as the Dutch disease effect (manifested through exchange rate appreciation), demand-side inflation, the “easy money syndrome”, and the emergence of “ghost towns” highlight potential adverse consequences (Thapa-Parajuli, 2013).

Amid these debates, there remains a clear research gap regarding whether rural Nepali households are utilizing their remittance income in a sustainable and effective manner. This study aims to address this gap by estimating consumption elasticities at a disaggregated level for rural Nepal, focusing on categories such as food, non-food items, health, education, alcohol, ritual-related expenses, and total consumption. This nuanced approach seeks to provide evidence on how remittances influence household consumption patterns and their implications for long-term sustainability. Moreover, the dataset used in this study was collected during a challenging period for Nepal’s economy, encompassing the post-earthquake years and a critical political transition. During this time, Nepal promulgated a new constitution, and the first elected governments were formed. The primary objective of the survey was to capture the vulnerability of rural Nepal, making it particularly relevant for examining the socioeconomic dynamics influenced by remittances.

Despite some concerns about the macroeconomic impacts of remittances, they have increasingly become a vital livelihood source for various economies, including Nepal. At the household level, there is an ongoing debate regarding the sustainability of remittance-funded consumption. However, the livelihood improvement potential of remittances cannot be ignored. This paper examines the relationship between remittance receipts and the consumption patterns of rural Nepali households during challenging times. By tracing consumption expenditures, this study seeks to shed light on the positive contribution of remittances to improving household welfare in rural Nepal. Our findings reveal that foreign remittances positively and significantly impact total consumption, food consumption, and non-food consumption in rural Nepal. Additionally, remittances significantly influence education and health expenditures, with the effect on education being more pronounced than that on health.

Interestingly, remittances do not significantly affect the consumption of unproductive ritual expenditures (temptation commodities that have low or no income elasticity to consumption) in Nepal. Contrary to common claims, rural households allocate minimal remittance income to rituals and other non-essential expenses such as alcohol and tobacco. Instead, remittance income positively affects household capital formation and food spending, while having a lesser impact on unproductive expenses. These findings challenge some existing research that takes a more negative view in this regard. Section 2 of this paper discusses the estimation strategy and the nature of the data. Section 3 presents the results and discussion, and Section 4 concludes the study. References and annexes are provided at the end, as usual.

2. Materials and Methods

2.1. Nature of the Data

As highlighted in the introduction, remittance receipts play a pivotal role in shaping household consumption patterns, particularly in rural Nepal where a significant proportion of the population depends on remittances for their livelihood. To explore the relationship between remittance inflows and consumption behavior, this paper utilizes data from the Nepal Household Risk and Vulnerability (NHRV) Survey Panel, conducted by the World Bank. This dataset covers rural areas across Nepal, offering a nationally representative sample that captures the diverse socioeconomic dynamics at the household level. The survey spans the years 2016 to 2018, encompassing 6000 households across 50 districts, ensuring broad geographic and demographic representation (World Bank, 2016).

Given the centrality of remittances to household welfare, especially during challenging economic times, these data provide an invaluable resource for analyzing consumption patterns. The inclusion of various variables related to household characteristics, income, and assets, alongside detailed expenditure data, enables an in-depth exploration of the role of remittances in enhancing household welfare. By utilizing this comprehensive dataset, we aim to provide a robust analysis of how remittances influence the consumption decisions of rural households in Nepal, with a particular focus on food, non-food, and human capital expenditures.

A total of 5648 households have been matched with all the variables under consideration. A summary of the variables is provided in

Table 1. Summary of the variables.

Headings	Subheadings	2016	2017	2018
Remittance receipt (in 000s)		53.89 (154.5)	58.64 (160.72)	60.02 (162.2)
Household size		4.88 (1.99)	4.48 (1.91)	4.46 (1.93)
Land size (ha)		0.45 (0.64)	0.45 (0.62)	0.46 (0.63)
Share of members older than 70 years		0.17 (0.44)	0.18 (0.46)	0.21 (0.48)
Wage income (in 000s)		61.18 (164.74)	71.06 (150.78)	81.08 (158.18)
Total assets (in 000s)		1229.46 (2555.78)	1367.28 (1857.39)	1474.1 (2506.4)
Household head’s years of schooling		4.65 (4.83)	4.62 (4.84)	4.57 (4.81)
House ownership (%)		97.8	99.2	98.8
Male-headed households (%)		81.2	77.2	76.0
Household having at least one shock (%)		16.1	25.1	13.1
Share of school-going children		0.28 (0.22)	0.28 (0.23)	0.27 (0.24)
Access to (in hours)	Secondary school	0.52 (0.59)	0.51(0.58)	0.52(1.00)
	Bank	1.79(3.52)	1.66(4.14)	1.55(2.71)
	Health post	0.67(1.39)	0.72 (4.28)	0.70 (4.10)
	Market	1.17 (3.76)	1.10 (5.36)	1.09 (5.36)
Expenditure (in 000)	Total	169.09 (98.09)	171.87 (166.11)	168.32 (108.43)
	Food	107.9 (56.95)	103.73 (54.65)	105.29 (50.99)
	Non-food	46.16 (45.65)	53.62 (140.07)	48.93 (52.04)
	Education	15.03 (27.58)	14.52 (27.38)	14.1 (58.57)
	Health	9.42 (32.62)	14.77 (44.94)	12.04 (50.59)
	Ritual activities	19.42 (45.86)	27.87 (94.32)	27.09 (64.37)
Marital status (%)	Married	88.8	88.1	87.0
	Unmarried	0.5	0.5	0.4
	Separate/divorce/wid.	10.7	11.5	12.6
Observation		5648	5648	5648

Note: Reported amounts are inflation adjusted; SE in parenthesis. Source: Author’s calculation.

. According to the data, the average total household consumption slightly increased in 2017 compared to 2016, but then decreased in 2018. A similar pattern is observed in the case of non-food consumption expenditure. In 2016, it was 169 thousand, which increased to 171 thousand in 2017 and then decreased in 2018. However, food consumption, asset acquisition, and wage income all show an upward trend. Furthermore, with the exception of ritual expenses, expenditures on education, non-food items, alcohol, and tobacco follow the same trend as total consumption. Notably, the average annual remittances from international migration grew from 61 thousand in 2016, showing an increasing pattern throughout the study period.

Additionally, household characteristics—particularly factors such as house ownership and the sex of the household head—have undergone changes over the years. For example, in 2016, the percentage of male-headed households was 19%, but this increased to 24% in 2018. Meanwhile, the average years of schooling for household heads remained relatively low at four years, as the study focuses on rural households in Nepal. When examining the ethnicity of Nepalese households, the dominant group is the Khas, which comprises one-third of the total households, followed by the Janajati.

2.2. Estimation Strategy

The data used in this study provide a comprehensive representation of rural households in Nepal, with a sufficiently large cross-section (n = 5648) over three years (T = 3). The dimensions of the panel data suggest the use of fixed-effects estimation (see Wooldridge (2010) for details). Specifically, we estimate total consumption and its disaggregated components in logarithmic form, controlling for various explanatory variables.

The utility of migration is maximized by enhancing the well-being of household members, as noted by (Lucas & Stark, 1985). This aligns with the Lewisian perspective, where migration to urban wage markets improves the productivity of remaining household members. Remittances serve as an altruistic contribution, improving household consumption and living standards. In rural Nepal, remittance receipts are a key source of prosperity. Additionally, the self-interest theory suggests that migrants remit to accumulate assets for their eventual return home (Fullenkamp et al., 2008). Whether driven by altruism or self-interest, remittances contribute to household disposable income, supporting consumption in food, non-food items, human capital formation, and sometimes even unproductive expenses like rituals.

Regardless of the motive to remit, consumption expenditures across various goods and services become a remittance-induced channel to prosperity. Hence, we regress household consumption expenditure $\left(E_{i,t}\right)$ on remittance receipts $\left(R_{i,t}\right)$, controlling for other relevant variables. The household consumption expenditure regression is as follows:

$$\begin{array}{c}\hfill ln{E}_{i,t}={\beta }_0+\delta R_{i,t}+X_{i,t}\beta +Z_i\lambda +T_t{\delta }_t+{ϵ}_{i,t}\end{array}$$

(1)

where $E_{i,t}$ is the expenditure of the ith household in year t, $R_{i,t}$ is the total remittance received, T is the year control variable, $X_{i,t}$ gives the household control variables, which include HH size, households assets, land size, people age 70 or more representing the dependent population at home, education of household head, and access to various services, and $Z_{i,t}$ represents the control for time-invariant fixed effects such as district. The letters ${\beta }_0$, $\delta$, $\beta$, $\lambda$, $\theta$, and ${\delta }_t$ represent respective parameters with ${ϵ}_{i,t}$ as an $i.i.d$ residual term.

2.3. Endogeneity and Identification Strategy

Endogeneity occurs when the regressor is correlated with the error term, leading to biased estimators. In our regression analysis, remittances may be associated with an unobserved variable, such as a household’s decision to send a member to the international labor market, which influences the household’s receipt of remittances. Consequently, if we estimate Equation (1) without correcting for endogeneity, the resulting estimators will be biased.

The Instrumental Variables (IVs) approach addresses the endogeneity problem, as Ullah et al. (2021) rightly pointed out and employed. According to the theory, the IVs should be correlated with the exogenous variables but not with the error terms of the model. Additionally, the IVs should not be associated with the outcome variable—in our case, consumption. Previous studies have used various instruments—e.g., Woodruff and Zenteno (2007) and Adams and Cuecuecha (2010) used the railroad network, while Bansak and Chezum (2009) used past literacy rates in similar estimations to ours.

Railroad connectivity is not applicable in the case of Nepal, and data on past literacy rates is not available in our dataset. While the Maoist insurgency and the resulting political unrest could serve as potential instruments, they occurred between 1996 and 2006, which does not align with our study’s time frame. Instead, we also use the migration network as an instrument, as employed by Mishra et al. (2022) and Mansuri (2006).

We construct our instrument, the migration network, using data from the National Census 2011, which meets both conditions for instrument validity. Since the migration network is not directly influenced by household consumption expenditure (our outcome variable), an increase in the migration network is expected to lead to higher migration rates and, consequently, increased household remittances (Massey & Espinosa, 1997).

While estimating the 2SLS equations for various consumption expenditures with foreign remittance receipts, incorporating multiple controls, we use the community-level migration network as our instrumental variable. The validity of this instrument is tested, with a summary of the results presented in

Table 2. Validity test of instrument.

Test Statistics Model	Weak Identification Test (Cragg–Donald Wald F Statistic)	Underidentification Test (Anderson Canon. Corr. LM Statistic)
Total consumption model	668.237 ***	642.956 ***
Food model	668.237 ***	642.956 ***
All model	668.237 ***	642.956 ***

Note: *** Model passed both tests. Source: Author’s calculation.

. The F-test statistic is well above the critical threshold of a weak identification test; we refer to the outline provided by Andrews and Stock (2005), indicating a strong positive correlation between foreign remittances and the migration network. Additionally, the under-identification test yields a highly significant result with a very low p-value, confirming that the structural model is identified correctly. Therefore, the instrument successfully passes both tests for validity.

3. Results and Discussion

3.1. Remittance and Households Consumption

We estimated 2SLS regressions on total household consumption expenditure, considering foreign remittances, household size, land ownership, the number of dependents over 70 years old, wage income, other durable assets (excluding land), and the education level of the household head. The foreign migration network variable is instrumented to foreign remittance. Additionally, access to roads, local markets, the nearest bank, health posts, and secondary schools were included as explanatory variables, as these factors also influence household expenditure in Nepal. We also recognize the district, year, male-headed households, and household head’s marital status to control their fixed effects on household consumption. Lastly, we also control household shocks, where the variable takes one when at least one shocks like death of the family member, fire, flood, critical disease of the family member and others are present, and zero otherwise.

We initially estimated a regression equation for total consumption using variables such as remittances, household size, land ownership, the proportion of dependent elderly family members, wage income, household assets, and the education level of the household head. These factors directly influence household consumption levels. Details on the variables’ measurement and scaling can be found in Appendix A. Additionally, we controlled for access to local amenities (roads, market centers, banks, health posts, and secondary schools), which can impact household consumption expenditure in rural Nepal. The results of this first regression are presented in

Table 3. Total consumption expenditure and remittance.

Variables:	Log Measure of Total Consumption Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.053 ***	0.038 ***	0.038 ***	0.047 ***	0.046 ***	0.046 ***
	(0.011)	(0.013)	(0.013)	(0.016)	(0.016)	(0.016)
Size hh	0.120 ***	0.130 ***	0.130 ***	0.124 ***	0.124 ***	0.124 ***
	(0.004)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Land size (ha)	0.093 ***	0.065 ***	0.065 ***	0.062 ***	0.061 ***	0.061 ***
	(0.010)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
Elderly (≥70 year old)	0.053	−0.083 *	−0.085 *	−0.106 **	−0.120 ***	−0.123 ***
	(0.055)	(0.049)	(0.050)	(0.047)	(0.040)	(0.040)
Wage income (log)	0.008 ***	0.007 ***	0.007 ***	0.007 **	0.007 ***	0.007 ***
	(0.002)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Assets (log) hh	0.146 ***	0.155 ***	0.153 ***	0.152 ***	0.151 ***	0.152 ***
	(0.010)	(0.007)	(0.007)	(0.008)	(0.008)	(0.008)
Education of hhh	0.027 ***	0.023 ***	0.023 ***	0.022 ***	0.022 ***	0.022 ***
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Fixed effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male hhh				Yes	Yes	Yes
Married hhh					Yes	Yes
Household shock						Yes
Access to roads, a market center, a bank, a health post, and a school are controlled.
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.1075	0.3617	0.3608	0.3037	0.3152	0.3195
Within R2		0.2879	0.2865	0.1999	0.2075	0.2083

Notes: Parentheses: Cluster SE. Significance: ***: 0.01, **: 0.05, *: 0.1. hh = Household; hhh = Household Head.

. The coefficients for the control variables are available in the

Table A2. Total consumption expenditure and remittance.

Variables:	Log Measure of Total Consumption Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.053 ***	0.038 ***	0.038 ***	0.047 ***	0.046 ***	0.046 ***
	(0.011)	(0.013)	(0.013)	(0.016)	(0.016)	(0.016)
Size hh	0.120 ***	0.130 ***	0.130 ***	0.124 ***	0.124 ***	0.124 ***
	(0.004)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Land size (ha)	0.093 ***	0.065 ***	0.065 ***	0.062 ***	0.061 ***	0.061 ***
	(0.010)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
Elderly (≥70 year old)	0.053	−0.083 *	−0.085 *	−0.106 **	−0.120 ***	−0.123 ***
	(0.055)	(0.049)	(0.050)	(0.047)	(0.040)	(0.040)
Wage income (log)	0.008 ***	0.007 ***	0.007 ***	0.007 **	0.007 ***	0.007 ***
	(0.002)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Assets (log) hh	0.146 ***	0.155 ***	0.153 ***	0.152 ***	0.151 ***	0.152 ***
	(0.010)	(0.007)	(0.007)	(0.008)	(0.008)	(0.008)
Education of hhh	0.027 ***	0.023 ***	0.023 ***	0.022 ***	0.022 ***	0.022 ***
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Monthly road access	−0.006 **	0.007 ***	0.007 ***	0.008 ***	0.008 ***	0.008 ***
	(0.003)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Distance to market	0.0007	−0.0008	−0.0008	−0.0009 *	−0.0009 *	−0.0009 *
	(0.001)	(0.0005)	(0.0005)	(0.0005)	(0.0005)	(0.0005)
Distance to bank	0.005 **	−0.002	−0.002	−0.002	−0.002	−0.002
	(0.002)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)
Distance to health post	−0.002 *	−0.002 ***	−0.002 ***	−0.002 ***	−0.002 ***	−0.002 ***
	(0.001)	(0.0005)	(0.0005)	(0.0005)	(0.0006)	(0.0006)
Distance to secondary school	0.026 *	−0.010	−0.010	−0.010	−0.010	−0.010
	(0.015)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)
Fixed-effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.10751	0.36168	0.36083	0.30369	0.31520	0.3195
Within R2		0.28787	0.28646	0.19999	0.20735	0.2083

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

,

Table A3. Full specification model.

Variables:	Total	Food	Non-Food	Education	Health	Ritual	Alco. & Tobacco
Variables
Remittance (log)	0.046 ***	0.033 **	0.044 **	0.265 ***	0.118 *	0.020	0.077
	(0.016)	(0.013)	(0.020)	(0.098)	(0.070)	(0.046)	(0.129)
Size hh	0.124 ***	0.125 ***	0.087 ***		0.251 ***	0.058 ***	0.155 ***
	(0.003)	(0.003)	(0.003)		(0.018)	(0.007)	(0.022)
Land size (ha)	0.061 ***	0.047 ***	0.089 ***	−0.099	0.092 *	0.106 ***	0.029
	(0.008)	(0.006)	(0.013)	(0.066)	(0.052)	(0.023)	(0.076)
Elderly (≥70 year old)	−0.123 ***	−0.094 ***	−0.163 ***		1.62 ***	−0.227 *	−1.05 ***
	(0.040)	(0.036)	(0.055)		(0.246)	(0.137)	(0.370)
Wage income(Log)	0.007 ***	0.005 **	0.012 ***	0.044 ***	0.020 *	0.008	0.060 ***
	(0.003)	(0.002)	(0.003)	(0.015)	(0.012)	(0.008)	(0.022)
Assets (log) hh	0.152 ***	0.118 ***	0.192 ***	0.219 ***	0.145 ***	0.241 ***	−0.181 ***
	(0.008)	(0.006)	(0.010)	(0.042)	(0.039)	(0.022)	(0.058)
Education of hhh	0.022 ***	0.014 ***	0.022 ***	0.107 ***	0.002	0.016 **	−0.144 ***
	(0.002)	(0.002)	(0.003)	(0.015)	(0.012)	(0.007)	(0.020)
Monthly road access	0.008 ***	0.007 ***	0.006 **	0.046 ***	0.024 **	−0.016 *	−0.064 ***
	(0.002)	(0.002)	(0.003)	(0.014)	(0.011)	(0.009)	(0.023)
Distance to market	−0.0009 *	−0.0006	−0.002 *	0.002	0.004	−0.003 ***	−0.012
	(0.0005)	(0.0004)	(0.001)	(0.006)	(0.003)	(0.001)	(0.010)
Distance to bank	−0.002	−0.001	−0.003	−0.009	−0.001	0.001	0.023
	(0.001)	(0.001)	(0.002)	(0.008)	(0.009)	(0.003)	(0.014)
Distance to health post	−0.002 ***	−0.001 ***	−0.002 *	−0.014	−0.002	−0.001	0.002
	(0.0006)	(0.0004)	(0.0009)	(0.010)	(0.003)	(0.001)	(0.008)
Distance to secondary school	−0.010	−0.010	−0.012 **	−0.111 **	0.017	−0.031 *	0.133
	(0.007)	(0.007)	(0.006)	(0.047)	(0.036)	(0.017)	(0.101)
Schoolgoers				2.77 ***
				(0.065)
Fixed effects
District	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Male HH head	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Marital status of head	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Household shock	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.31951	0.35618	0.25782	0.52224	0.30608	0.15352	0.15907
Within R2	0.20832	0.25206	0.12204	0.50100	0.00114	0.03716	0.04013

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

,

Table A4. Remittance and food expenditure in Nepal (full model).

Variables:	Log of Food Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.056 ***	0.026 **	0.026 **	0.034 **	0.033 **	0.033 **
	(0.011)	(0.011)	(0.010)	(0.013)	(0.013)	(0.013)
Size hh	0.125 ***	0.131 ***	0.131 ***	0.125 ***	0.125 ***	0.125 ***
	(0.004)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Land size (ha)	0.079 ***	0.051 ***	0.050 ***	0.048 ***	0.047 ***	0.047 ***
	(0.012)	(0.006)	(0.006)	(0.006)	(0.006)	(0.006)
Elderly (≥70 year old)	0.122 **	−0.060	−0.063	−0.082 **	−0.091 **	−0.094 ***
	(0.053)	(0.043)	(0.043)	(0.041)	(0.035)	(0.036)
Wage income (log)	0.008 ***	0.005 **	0.005 **	0.005 **	0.005 **	0.005 **
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Assets (log) hh	0.101 ***	0.119 ***	0.119 ***	0.117 ***	0.117 ***	0.118 ***
	(0.009)	(0.006)	(0.005)	(0.006)	(0.006)	(0.006)
Education of hhh	0.021 ***	0.014 ***	0.014 ***	0.013 ***	0.014 ***	0.014 ***
	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Monthly road access	−0.008 ***	0.006 ***	0.006 ***	0.007 ***	0.007 ***	0.007 ***
	(0.003)	(0.002)	(0.002)	(0.002)	(0.002)	(0.002)
Distance to market	0.001	−0.0005	−0.0005	−0.0005	−0.0005	−0.0006
	(0.001)	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)
Distance to bank	0.005 **	−0.002	−0.002	−0.001	−0.001	−0.001
	(0.002)	(0.001)	(0.002)	(0.001)	(0.001)	(0.001)
Distance to health post	−0.001	−0.001 ***	−0.001 ***	−0.001 ***	−0.001 ***	−0.001 ***
	(0.001)	(0.0005)	(0.0005)	(0.0005)	(0.0004)	(0.0004)
Distance to secondary school	0.030 **	−0.009	−0.009	−0.010	−0.010	−0.010
	(0.015)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)
Fixed effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.01224	0.37830	0.38069	0.34540	0.35137	0.35618
Within R2		0.30939	0.31041	0.24784	0.25089	0.25206

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

,

Table A5. Remittance and non-food expenditure (full model).

Variables:	Log of Non-Food Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.044 ***	0.029 *	0.030 *	0.045 **	0.044 **	0.044 **
	(0.013)	(0.017)	(0.017)	(0.021)	(0.020)	(0.020)
Size hh	0.082 ***	0.094 ***	0.098 ***	0.087 ***	0.087 ***	0.087 ***
	(0.004)	(0.004)	(0.003)	(0.003)	(0.003)	(0.003)
Land size (ha)	0.131 ***	0.095 ***	0.095 ***	0.090 ***	0.089 ***	0.089 ***
	(0.014)	(0.013)	(0.013)	(0.013)	(0.013)	(0.013)
Elderly (≥70 year old)	0.033	−0.104	−0.110	−0.145 **	−0.160 ***	−0.163 ***
Wage income (log)	0.013 ***	0.012 ***	0.011 ***	0.012 ***	0.012 ***	0.012 ***
	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Assets (log) hh	0.206 ***	0.204 ***	0.195 ***	0.192 ***	0.191 ***	0.192 ***
	(0.013)	(0.010)	(0.010)	(0.010)	(0.010)	(0.010)
Education of hhh	0.025 ***	0.023 ***	0.023 ***	0.021 ***	0.022 ***	0.022 ***
	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Monthly road access	−0.006	0.005 *	0.005 *	0.006 **	0.006 **	0.006 **
	(0.004)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Distance to market	$6.2\times {10}^{-5}$	−0.002 *	−0.002 *	−0.002 **	−0.002 **	−0.002 *
	(0.070)	(0.067)	(0.068)	(0.062)	(0.054)	(0.055)
	(0.002)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)
Distance to bank	0.004	−0.004	−0.004	−0.003	−0.003	−0.003
	(0.003)	(0.003)	(0.002)	(0.002)	(0.002)	(0.002)
Distance to health post	−0.002	−0.002 **	−0.002 ***	−0.002 **	−0.002 **	−0.002 *
	(0.001)	(0.0007)	(0.0007)	(0.0008)	(0.0009)	(0.0009)
Distance to secondary school	0.028	−0.010 *	−0.011 *	−0.012 **	−0.012 **	−0.012 **
	(0.022)	(0.006)	(0.006)	(0.006)	(0.006)	(0.006)
Fixed effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.11952	0.28434	0.28548	0.24685	0.25480	0.25782
Within R2		0.19736	0.19005	0.11549	0.12131	0.12204

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

,

Table A6. Remittance and education expenditure (full model).

Variables:	Log of Education Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.172 ***	0.255 ***	0.255 ***	0.271 ***	0.265 ***	0.265 ***
	(0.053)	(0.080)	(0.080)	(0.103)	(0.098)	(0.098)
School goers	2.72 ***	2.76 ***	2.76 ***	2.76 ***	2.77 ***	2.77 ***
	(0.065)	(0.064)	(0.063)	(0.064)	(0.065)	(0.065)
Land size (ha)	−0.085	−0.086	−0.085	−0.093	−0.099	−0.099
	(0.069)	(0.067)	(0.067)	(0.066)	(0.066)	(0.066)
Wage income (Log)	0.032 ***	0.045 ***	0.045 ***	0.045 ***	0.044 ***	0.044 ***
	(0.011)	(0.015)	(0.016)	(0.016)	(0.015)	(0.015)
Assets (log) hh	0.320 ***	0.233 ***	0.228 ***	0.222 ***	0.218 ***	0.219 ***
	(0.040)	(0.038)	(0.038)	(0.042)	(0.042)	(0.042)
Education of hhh	0.089 ***	0.104 ***	0.104 ***	0.103 ***	0.107 ***	0.107 ***
	(0.012)	(0.015)	(0.015)	(0.013)	(0.015)	(0.015)
Monthly road access	0.076 ***	0.045 ***	0.045 ***	0.047 ***	0.046 ***	0.046 ***
	(0.012)	(0.013)	(0.013)	(0.015)	(0.014)	(0.014)
Distance to market	0.002	0.003	0.003	0.003	0.002	0.002
	(0.007)	(0.006)	(0.006)	(0.006)	(0.006)	(0.006)
Distance to bank	−0.016 *	−0.010	−0.010	−0.010	−0.009	−0.009
	(0.008)	(0.007)	(0.007)	(0.008)	(0.008)	(0.008)
Distance to health post	−0.010	−0.014	−0.015	−0.014	−0.014	−0.014
	(0.009)	(0.010)	(0.010)	(0.010)	(0.010)	(0.010)
Distance to secondary school	−0.091 **	−0.110 **	−0.110 **	−0.111 **	−0.111 **	−0.111 **
	(0.042)	(0.047)	(0.047)	(0.048)	(0.047)	(0.047)
Fixed effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.55289	0.52694	0.52676	0.51769	0.52226	0.52224
Within R2		0.51435	0.51412	0.50265	0.50102	0.50100

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

,

Table A7. Remittance and health expenditure (full model).

Variables:	Log of Health Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.027	0.049	0.095	0.122	0.117	0.118 *
	(0.059)	(0.065)	(0.063)	(0.078)	(0.075)	(0.070)
Size hh	0.164 ***	0.171 ***	0.281 ***	0.262 ***	0.262 ***	0.251 ***
	(0.021)	(0.020)	(0.020)	(0.019)	(0.019)	(0.018)
Land size (ha)	0.041	0.104 *	0.102 *	0.092	0.089	0.092 *
	(0.057)	(0.056)	(0.056)	(0.057)	(0.056)	(0.052)
Elderly (≥70 year old)	1.86 ***	1.99 ***	1.82 ***	1.76 ***	1.70 ***	1.62 ***
	(0.303)	(0.304)	(0.304)	(0.286)	(0.263)	(0.246)
Wage income (Log)	0.028 **	0.029 **	0.019	0.020	0.018	0.020 *
	(0.013)	(0.013)	(0.013)	(0.014)	(0.013)	(0.012)
Assets (log) hh	0.359 ***	0.391 ***	0.119 ***	0.113 ***	0.110 ***	0.145 ***
	(0.049)	(0.045)	(0.039)	(0.040)	(0.041)	(0.039)
Education of hhh	−0.025 *	−0.020	0.004	−0.0007	0.003	0.002
	(0.013)	(0.013)	(0.013)	(0.011)	(0.013)	(0.012)
Monthly road access	0.032 **	0.036 **	0.022 *	0.025 **	0.024 **	0.024 **
	(0.016)	(0.015)	(0.012)	(0.012)	(0.012)	(0.011)
Distance to market	0.003	0.004	0.004	0.004	0.004	0.004
	(0.006)	(0.006)	(0.005)	(0.005)	(0.005)	(0.003)
Distance to bank	−0.011	−0.011	−0.004	−0.003	−0.003	−0.001
	(0.013)	(0.012)	(0.009)	(0.009)	(0.009)	(0.009)
Distance to health post	0.002	0.0002	−0.004	−0.004	−0.004	−0.002
	(0.004)	(0.004)	(0.003)	(0.003)	(0.003)	(0.003)
Distance to secondary school	0.091 *	0.050	0.018	0.017	0.017	0.017
	(0.047)	(0.041)	(0.040)	(0.039)	(0.039)	(0.036)
Fixed effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.02692	0.06207	0.22481	0.21720	0.21994	0.30608
Within R2		0.02477	0.01230	0.00037	0.00284	0.00114

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

,

Table A8. Remittance and alcohol and tobacco consumption (full model).

Dependent Variable:	Log of Alcohol and Tobacco Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Variables
Remittance (log)	−0.087	−0.100	−0.100	0.085	0.077	0.077
	(0.073)	(0.120)	(0.119)	(0.133)	(0.129)	(0.129)
Size hh	0.238 ***	0.284 ***	0.283 ***	0.152 ***	0.155 ***	0.155 ***
	(0.026)	(0.027)	(0.026)	(0.022)	(0.022)	(0.022)
Land size (ha)	0.085	0.104	0.106	0.040	0.029	0.029
	(0.089)	(0.083)	(0.083)	(0.077)	(0.076)	(0.076)
Elderly (≥70 year old)	−0.464	−0.529	−0.510	−0.944 **	−1.05 ***	−1.05 ***
	(0.411)	(0.481)	(0.486)	(0.419)	(0.370)	(0.370)
Wage income (log)	0.063 ***	0.056 **	0.057 **	0.063 ***	0.060 ***	0.060 ***
	(0.017)	(0.024)	(0.025)	(0.023)	(0.022)	(0.022)
Assets (log) hh	−0.235 ***	−0.142 **	−0.139 **	−0.177 ***	−0.181 ***	−0.181 ***
	(0.060)	(0.057)	(0.056)	(0.056)	(0.058)	(0.058)
Education of hhh	−0.117 ***	-0.119 ***	−0.119 ***	−0.149 ***	−0.144 ***	−0.144 ***
	(0.016)	(0.022)	(0.021)	(0.017)	(0.020)	(0.020)
Monthly road access	−0.112 ***	−0.081 ***	−0.081 ***	−0.063 ***	−0.064 ***	−0.064 ***
	(0.024)	(0.025)	(0.025)	(0.023)	(0.023)	(0.023)
Distance to market	−0.005	−0.011	−0.011	−0.012	−0.012	−0.012
	(0.009)	(0.009)	(0.009)	(0.010)	(0.010)	(0.010)
Distance to bank	0.038	0.017	0.017	0.023	0.023	0.023
	(0.023)	(0.015)	(0.015)	(0.014)	(0.014)	(0.014)
Distance to health post	0.0005	−0.002	−0.002	0.002	0.002	0.002
	(0.008)	(0.007)	(0.007)	(0.008)	(0.008)	(0.008)
Distance to secondary school	0.199	0.142	0.144	0.135	0.133	0.133
	(0.130)	(0.112)	(0.111)	(0.103)	(0.101)	(0.101)
Fixed-effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.07566	0.11051	0.11101	0.15626	0.15902	0.15907
Within R2		0.05945	0.05895	0.03943	0.04009	0.04013

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

and

Table A9. Remittance and ritual expenditure (full model).

Variables:	Log of Ritual Expenditure
	(1)	(2)	(3)	(4)	(5)	(6)
Remittance (log)	0.058 *	0.016	0.018	0.023	0.020	0.020
	(0.035)	(0.040)	(0.040)	(0.048)	(0.046)	(0.046)
Size hh	0.038 ***	0.054 ***	0.063 ***	0.059 ***	0.059 ***	0.058 ***
	(0.009)	(0.009)	(0.009)	(0.007)	(0.007)	(0.007)
Land size (ha)	0.174 ***	0.113 ***	0.110 ***	0.108 ***	0.106 ***	0.106 ***
	(0.035)	(0.024)	(0.024)	(0.023)	(0.023)	(0.023)
Elderly (≥70 year old)	−0.017	−0.126	−0.172	−0.184	−0.219	−0.227 *
	(0.166)	(0.169)	(0.170)	(0.154)	(0.138)	(0.137)
Wage income (log)	0.020 ***	0.011	0.009	0.009	0.008	0.008
	(0.007)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
Assets (Log) hh	0.290 ***	0.262 ***	0.241 ***	0.240 ***	0.238 ***	0.241 ***
	(0.028)	(0.022)	(0.021)	(0.022)	(0.022)	(0.022)
Education of hhh	0.018 **	0.013 *	0.015 **	0.014 **	0.016 **	0.016 **
	(0.007)	(0.008)	(0.007)	(0.006)	(0.007)	(0.007)
Monthly road access	0.003	−0.015	−0.016 *	−0.015 *	−0.016 *	−0.016 *
	(0.009)	(0.010)	(0.009)	(0.009)	(0.009)	(0.009)
Distance to market	0.002	−0.003 ***	−0.003 ***	−0.003 ***	−0.003 ***	−0.003 ***
	(0.003)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)
Distance to bank	−0.005	−0.0008	0.0007	0.0009	0.001	0.001
	(0.005)	(0.003)	(0.003)	(0.003)	(0.003)	(0.003)
Distance to health post	−0.001	−0.001	−0.002	−0.001	−0.002	−0.001
	(0.002)	(0.002)	(0.001)	(0.001)	(0.001)	(0.001)
Distance to secondary school	0.056	−0.026	−0.031 *	−0.031 *	−0.031 *	−0.031 *
	(0.039)	(0.018)	(0.017)	(0.017)	(0.017)	(0.017)
Fixed-effects
District		Yes	Yes	Yes	Yes	Yes
Year			Yes	Yes	Yes	Yes
Male HH head				Yes	Yes	Yes
Marital status of head					Yes	Yes
Household shock						Yes
Fit statistics
Observations	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.03111	0.13000	0.14786	0.14737	0.14937	0.15352
Within R2		0.03934	0.03765	0.03535	0.03660	0.03716

Notes: (a) Cluster SE in parentheses. (b) Significance: ***: 0.01, **: 0.05, *: 0.1. (c) hh = Household; hhh = Household Head.

in Appendix A. We report the coefficients of our interest variables only.

We controlled for district fixed effects with alternative specification, and the summary of the coefficients is given in the second regression of Table 3. Additionally, we successively controlled for year, gender of the household head, marital status of the household head, and the prevalence of household shocks. Regressions with these controls are summarized successively in Models (3) through (6) in the same Table. Among these controls, household shocks are particularly noteworthy, which capture events such as the death of a family member, critical illness of dependents, and natural disasters like fires or floods. The household is considered to have experienced a shock if it faced at least one of these events, and shock-free otherwise (details on measurement and scalling of the variables are given in

Table A1. Definition of variables.

Variables	Construct	Source
Total consumption	Sum of food and non-food, education	NHRVS: section (2–6)
Education expense	Total expenses for education	NHRVS: section (2)
Health expense	Total expenses for health services	NHRVS: section (3)
Alcohol–tobacco	Total consumption of alcohol and tobacco	NHRVS: section (6)
Household assets	Sum of financial assets including value of stocks, deposits, and net lending; physical assets including HH inventory and house value; agriculture assets comprising value of livestock and agri. inventory; and business assets	NHRVS: section (4, 6, 9, 12)
Remittance income	Total income received by household from international migration in a year	NHRVS: section (11)
Size of land	Total size of land in hector	NHRVS: section (9)
Wage income	Total income in cash and kind received by households	NHRVS: section (8)
Household size	Size of family member	NHRVS: section (1)
HH size aged ≥70	Ratio of total hh member with age 70 and more to total HH size	NHRVS: section (1)
Male	Male = 1 and Female = 0
Education	Education status (illiterate, below primary, primary, tenth grade, secondary, bachelor, masters, and above—0, 1, 2, 3, 4, 5, and 6, respectively)	NHRVS: section (1)
Services access	School, road, market (in hours)	NHRVS: section (4)
Access to road	Number of accessible roads each month
Marital status	Married, and unmarried, divorce/separate/widow—codes: 0, 1, and 2, respectively
Shocks	Households shocks including death of family members, fire, disease of HH members, etc.	NHRVS: section (15)
CPI	We used regional and national CPI values from 2015, 2016, and 2017 to correct regional variation due to inflation	NRB: datasets
Migration network	Ratio of migration size to HH size, constructed using national census of 2011	Census 2011: absentees
Schoolgoers	Ratio of currently enrolled HH members to total household size	NHRVS: section (2)

Note: Expenses and income measured during a year; NHRVS = Nepal Household Risk and Vulnerability Survey 2016–2018; NRB = Nepal Rastra Bank. HH = Household.

in Appendix A).

The coefficients of determination, which are at an acceptable level in all the models, both overall and within, remain consistent across Models (1) through (6) for our sufficiently large sample (n = 16,944) observations. We claim that we estimated the best available IV regression estimates, as we control for five key access-related variables in all models, along with an additional five fixed effects step by step. We summarized the coefficients without these ten control coefficients to produce Table 3 and

Table 4. Consumption expenditures and foreign remittance.

	Log Measure of Consumption Expenditures
Variables:	Total	Food	Non-Food	Edu.	Health	Ritual	Alcohol
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Remittance (log)	0.046 ***	0.033 **	0.044 **	0.265 ***	0.118 *	0.020	0.077
	(0.016)	(0.013)	(0.020)	(0.098)	(0.070)	(0.046)	(0.129)
Size hh	0.124 ***	0.125 ***	0.087 ***		0.251 ***	0.058 ***	0.155 ***
	(0.003)	(0.003)	(0.003)		(0.018)	(0.007)	(0.022)
Land size (ha)	0.061 ***	0.047 ***	0.089 ***	-0.099	0.092 *	0.106 ***	0.029
	(0.008)	(0.006)	(0.013)	(0.066)	(0.052)	(0.023)	(0.076)
Elderly (≥70 year old)	−0.123 ***	−0.094 ***	−0.163 ***		1.62 ***	−0.227 *	−1.05 ***
	(0.040)	(0.036)	(0.055)		(0.246)	(0.137)	(0.370)
Wage income (log)	0.007 ***	0.005 **	0.012 ***	0.044 ***	0.020 *	0.008	0.060 ***
	(0.003)	(0.002)	(0.003)	(0.015)	(0.012)	(0.008)	(0.022)
Assets (log) hh	0.152 ***	0.118 ***	0.192 ***	0.219 ***	0.145 ***	0.241 ***	−0.181 ***
	(0.008)	(0.006)	(0.010)	(0.042)	(0.039)	(0.022)	(0.058)
Education of hhh	0.022 ***	0.014 ***	0.022 ***	0.107 ***	0.002	0.016 **	−0.144 ***
	(0.002)	(0.002)	(0.003)	(0.015)	(0.012)	(0.007)	(0.020)
School goers				2.77 ***
				(0.065)
District, year, male hhh, maried hhh, and household shock fixed effects are controlled. We also control access to roads, a market centre, a bank, a health post, and a school (see Appendix A).
Observations	16,944	16,944	16,944	16,944	16,944	16,944	16,944
R2	0.3195	0.3562	0.2578	0.5222	0.3061	0.1535	0.1591
Within R2	0.2083	0.2521	0.1220	0.5010	0.0011	0.0372	0.0401

Notes: Parenthesis: Cluster SE. Significance: ***: 0.01, **: 0.05, *: 0.1. hh = Household; hhh = Household Head.

and present the full information in Table A2 and Table A3 in Appendix A.

The coefficient of our primary interest, the remittance, is consistently positive and statistically significant even at the 1% level of significance across all six specifications. This indicates that remittance receipt significantly and positively influences total household consumption in rural Nepal, holding other factors constant. The remittance elasticity (0.046) coefficient is 15% larger than wage income elasticity (0.007), indicating the level of influence of remittance on household total consumption. Similarly, household size also positively affects total household consumption expenditure. A large-sized household not only demands more consumption but also has the potential to generate more resources, thus increasing total consumption. A similar relationship is observed for household land size and other household assets, where wealthier households can afford and might have spent more, resulting in higher consumption levels. As expected, all coefficients for these variables are positive and statistically significant, aligning with theoretical predictions.

The coefficient for wage income is positive and statistically significant, indicating that higher wage income leads to increased consumption, which is also true in rural Nepal. In contrast, the share of elderly individuals (those over 70 years old) in a household is negatively associated with total household consumption. The presence of elderly members significantly reduces household expenditure, possibly due to their lower consumption habits, reliance on locally produced goods, or even their contribution to household income rather than consumption. However, we are constrained by data limitations in exploring these details further. The share of school-going children in a household is positively and significantly associated with total consumption, unlike the pattern observed with elderly members. As the number of school-going children in a household increases, so do household expenditures. These dependent family members, in general, do not contribute to household income or chores.

The regression coefficients for total household consumption, which includes spending on food, non-food items, health, education, and rituals are summarized in Table 3. The results show that foreign remittance has a positive and significant effect on overall consumption. However, remittance may have had differing impacts across the various categories of disaggregated consumption. To explore this, we estimated separate regressions for food, non-food, education, health, rituals, and alcohol-related expenditures, in addition to total consumption. The results of these regressions are summarized in Table 4.

Regression (1) in Table 4 is replica of Regression (6) in Table 3, in which we included all five access-related controls and another set of five fixed effects. We estimated six other regression models for each disaggregate level expenditure category: food, non-food, education, health, ritual-related, and alcohol–tobacco expenditure. Similar to total consumption and summarized in Table 3, we systematically estimated the alternative regressions for each sub-group. Keeping the details in the annexe, we present the summary of the full models in (2) through (7) in Table 4. A full estimation of each of Regressions (2) through (7) is given in Table A4, Table A5, Table A6, Table A7, Table A8 and Table A9 in Appendix A. The regression for education differs slightly from the others; it does not include the elderly dependency variable and substitutes household size with the ratio of school-going children to household size to make the regression more realistic.

According to the regression results, remittances positively and statistically significantly affect food and non-food consumption. The elasticity coefficient for remittances is higher for non-food consumption than for food consumption. This suggests that remittances enhance food consumption, improving food intake and nutrition in rural Nepal. Additionally, remittances boost non-food consumption, particularly for household durables and other welfare-enhancing items such as education and healthcare. Aside from the share of elderly dependent family members, all other coefficients in both the food and non-food consumption models are positive and significant. Both models exhibit an acceptable level of goodness of fit.

Our primary interest among non-food consumption expenditures is in education and health. The remittance elasticity coefficient for education is positive and statistically significant, as shown in Model (4) of Table 4. A 1% increase in foreign remittances results in a 0.27% increase in household education expenses, which is a substantial coefficient. In rural Nepal, where public schools are more common and less costly, a larger portion of education expenses is allocated to non-fee-related items contributing to educational attainment and quality. Similarly, the remittance elasticity for health expenditure is also positive and statistically significant and relatively high in Model (5) of Table 4. All other coefficients in the health model are positive and statistically significant. However, the coefficient for the education level of the household head is insignificant, meaning that the head’s education hardly matters for the level of health expenditure, a typical health financing characteristic. Logically, higher health expenditures are associated with a more significant share of elderly dependent family members, leading to a significantly higher coefficient. Therefore, among non-food consumption expenditures, household investments in human capital improvements—such as health and education—are notably influenced by foreign remittances in rural Nepal.

We also analyzed household expenses related to ritual activities and alcohol and tobacco consumption, with the results shown in Models (6) and (7) of Table 4. These expenditures might have contributed to the social status of the family level satisfaction and potentially enhanced local economic activities considered unproductive, with remittance being blamed for catering to consumption of such goods and services. However, our findings indicate that the elasticities of remittances with respect to alcohol and ritual expenses are statistically insignificant.

Households with elderly members incur fewer ritual-related expenses, likely due to the absence of recent deaths and associated rituals. Conversely, wealthier households with more assets and higher education spend significantly more on ritual-related items. Interestingly, wage income is less likely to be spent on rituals, possibly because hard-earned money is valued more. The results regarding alcohol consumption reveal that larger families with wage income are more likely to consume alcohol, as indicated by the positive and significant coefficients. Higher household education levels, greater asset ownership, and the presence of elderly dependents are associated with reduced alcohol consumption. Senior citizens often act as moral guardians against alcohol consumption in rural Nepal. Additionally, more assets minimize the consumption of tempting goods, and higher education levels foster a better understanding of sustainable consumption, a healthier lifestyle, and improved financial literacy. Land size does not appear to affect alcohol consumption in rural Nepal. In conclusion, remittances do not induce spending on rituals and alcohol, ceteris paribus.

3.2. Discussion

Nepal is expected to graduate from an underdeveloped economy to a developing country in 2026. Despite acceptable social and environmental parameters, Nepal’s lower national income prevents the country’s LDC graduation. Even with shortcomings in one parameter, one can qualify that the Nepali government has opted to wait until 2026 to ensure the necessary preparations for a smooth transition.

The country’s overall growth in the last five years has been sluggish, hovering around 4–5%. The trade deficit remains substantial, with the export-to-import ratio being one to nine, and most of the trade deficit is with India. Remittances are crucial in offsetting the import imbalance despite weak export performance. Both remittance inflows and tourism contribute significantly to maintaining the country’s foreign reserves. Development financing increasingly relies on deficit financing, with domestic and foreign deficits growing over time. There is fear-mongering associated with graduation claiming that instead of the country being branded as a progressing economy, sources of concessional development assistance might be curtailed.

In this context, one cannot forget that remittances play a vital role in improving health and education at the household level despite the poor performance of government development delivery. The increasing trend in internal and foreign migration rapidly changes the economic landscape, particularly in rural areas. The analysis of foreign remittances and their impact on recipient economies offers two contrasting perspectives: the livelihood approach and the macroeconomic angle. From a livelihood perspective, remittances are vital for improving recipients’ livelihoods. They serve as informal insurance for needy households and help reduce poverty. This perspective aligns with the Lewisian theory of unlimited labor supply, where migration to urban areas or regions with higher wages benefits migrants and their families who remain at home. However, this approach often overlooks direct and indirect remittance costs. In search of higher wages, rural Nepali people migrate to OECD countries, the Gulf region, Asia, and India, depending on their ability to finance migration costs (Kshetri & Thapa-Parajuli, 2022).

Foreign remittances are international transfer payments. The transfer payment multiplier is just unity; one extra unit injection generates the same one, and there is no value addition in the economy. This is a skeptical argument about the long-term sustainability of remittances and their potential to contribute to sustained economic growth and development. The effects of remittance-induced Dutch disease and the laxity of policy to improve the investment climate in Nepal (Sapkota, 2013) and remittances are used in consumption but not investment (Thapa-Parajuli, 2013). These differing views have proponents and critics, and the debate continues. This paper focuses on examining short-term consumption patterns and their association with remittances in rural households in Nepal.

The social costs associated with remittances, such as the absence of key family members and the unsustainable use of resources at the national level, are increasingly being discussed in the literature. However, the evidence from rural Nepal suggests a less pessimistic outlook. According to the results in Table 4, foreign remittances have a significant positive impact on the welfare of rural households. They enhance various aspects of consumption, including food, non-food items, health, and education expenses. These findings are particularly noteworthy given that remittances are a major income source for rural households in Nepal, a country that ranks tenth globally in terms of the foreign remittance-to-GDP ratio (World Bank, 2023a). When consumption increases, aggregate demand might play a catalyst role in further economic expansion. However, the ground reality might differ, and some homegrown policies might help to channel demand-driven remittance shocks to the productive supply chain.

Our results indicate that rural households primarily use their remittance income to enhance human capital, followed by expenditures on non-food and food items, with the highest elasticity observed in these categories. This finding is consistent with Mishra et al. (2022), who similarly found that remittances significantly and positively affect education and food consumption among Nepali households. Additionally, Bansak et al. (2015) and Raut and Tanaka (2018) draw similar conclusions, emphasizing the positive impact of remittances on child education within Nepali households. Furthermore, remittances also help finance health expenses for rural households, aligning with the results found by Kapri and Jha (2020). Therefore, positive and significant improvements in health and education expenditure might have improved the human capital formation in rural Nepal, which might have helped not only gain sustainable development goals but also complemented the means of production, and hence growth, as Mishra et al. (2022) reveal in the case of Nepal and Wang et al. (2019) in the case of Kyrgyzstan.

One of the key findings of this paper is the insignificant association between remittances and spending on temptation goods, generally less productive expenditure. Although the coefficients are positive, they are not statistically significant, likely due to the nature of these products. A similar trend is observed for ritual-related expenses and social status variables. With higher income, these commodities are generally expected to be consumed more. While such expenditures are often seen as unproductive, they may enhance family satisfaction and social status, and potentially stimulate local economic activities. This could be a distinct phenomenon in rural Nepal, although the extent to which remittances significantly improved unproductive expenses like wedding expenses, ritual expenses, and flamboyant gifts in the case of Kyrgyzstan remains unclear (Reeves, 2012; Wang et al., 2019). Remittances are sometimes blamed for driving consumption of these goods, but our results show that the elasticities of remittances to alcohol and ritual expenses are statistically insignificant. One possible argument would be that the presence of a person older than 70 years old means no recent deaths in a household. And, marriage-age household members in general are absent from home for foreign employment. Wedding expenses and ritual expenses are major headings to scale up ritual expenses in rural Nepal. Moreover, the absence of major household members might curtail festival expenses, too.

Households with elderly members tend to spend less on rituals, likely due to the absence of recent deaths. On the other hand, wealthier households with more assets and higher education levels spend significantly more on rituals. Interestingly, wage income is less likely to be spent on rituals, possibly because hard-earned money is more carefully allocated. Larger families with wage income are more likely to consume alcohol, while higher education, greater asset ownership, and the presence of elderly members reduce alcohol consumption; this finding is pretty similar to Mishra et al. (2022). In rural Nepal, senior citizens often act as moral guardians against alcohol consumption, and higher education encourages more sustainable consumption and healthier lifestyles. Land size does not influence alcohol consumption.

Overall, remittances do not significantly drive spending on rituals or alcohol. Household assets and characteristics like family size better explain ritual expenditures. Similarly, remittances have no significant impact on alcohol and tobacco consumption, likely due to the habitual nature of these products. Households with more education are also more likely to discourage alcohol and tobacco use.

4. Conclusions and Recommendations

This study employs Two-Stage Least Squares (2SLS) regression to analyze the impact of foreign remittances on total household consumption expenditures in rural Nepal. The analysis includes various factors such as household size, land ownership, elderly dependents, wage income, other durable assets, and the education level of the household head. Additionally, the study incorporates instrumental variables for remittances, such as access to infrastructure and local amenities, including roads, markets, banks, health posts, and schools, which are relevant for household expenditures.

Initial results show that foreign remittances significantly and positively affect total household consumption. This positive effect is consistent across multiple specifications and controls for district, year, gender, marital status of the household head, and household shocks (e.g., death, illness, natural disasters). The study also finds that larger household size and land ownership are positively associated with increased consumption. In contrast, the presence of elderly members tends to reduce total consumption, potentially due to their lower consumption needs or their contributions to household income.

Further disaggregation of consumption into categories such as food, non-food, education, health, rituals, and alcohol reveals nuanced effects of remittances. Specifically, remittances significantly boost food and non-food consumption, with a higher elasticity for non-food items. Remittances notably influence investment in education and health. However, expenditures on rituals and alcohol are not significantly affected by remittances, indicating that these transfers are not driving consumption in these areas. Overall, remittances enhance consumption in critical areas such as education and health while having negligible effects on ritual and alcohol-related expenditures. Therefore, remittance receipts help the consumption of food items and improve health and education, not necessarily catering to the unproductive use of remittance receipts like ritual-related expenses and temptation goods consumption like alcohol and tobacco; this is an optimistic scenario.

Governments, particularly local governments, can use these findings as a guideline to improve policy effectiveness. Nepali rural households are utilizing hard-earned remittances in sustainable ways, contributing to the achievement of the Sustainable Development Goals (SDGs). Enhancing infrastructure and access to essential services such as roads, markets, banks, health posts, and schools can significantly improve social welfare in rural Nepal, as households are both willing and able to pay for these services sustainably.

By improving such infrastructure, local governments can reduce the need for transfer payments and limit direct contributions to these programs. However, this does not imply that promoting remittances should be the focus. Instead, living arrangements with elderly family members seem to help reduce unnecessary expenses, such as spending on alcohol and rituals. Therefore, providing urban-like amenities in and near rural areas could encourage parents and children to live with grandparents, fostering a more sustainable use of remittance income.

Expanding this analysis to include urban areas at the national level could offer a broader understanding of remittance impacts. Further disaggregation of findings, focusing on calorie intake and SDG-specific outcomes, may uncover deeper insights into food security and sustainable development.