# Local infrastructure, rural households' resilience capacity and poverty: evidence from panel data for Southeast Asia

Tim Hartwig (Leibniz University Hannover, Hannover, Germany)
Trung Thanh Nguyen (Leibniz University Hannover, Hannover, Germany)

ISSN: 1859-0020

Article publication date: 27 December 2022

112

## Abstract

### Purpose

The authors examine the association between infrastructure and a household's resilience capacity against shocks and the impacts of a household's resilience capacity on household consumption and poverty.

### Design/methodology/approach

The authors use panel data (collected in 2010, 2013 and 2016) from 1,698 households in Thailand and 1,701 households in Vietnam and employ an instrumental variable approach.

### Findings

The authors find that transportation and information and communication technology (ICT) infrastructure help improve households' absorptive capacity in coping with shocks. Furthermore, this capacity can prevent households from reducing consumption and falling into poverty.

### Practical implications

Rural development policies should attend to transportation and ICT infrastructure.

### Originality/value

The authors establish empirical evidence on the association between infrastructure and a household's resilience capacity and the impact of resilience capacity on poverty.

## Citation

Hartwig, T. and Nguyen, T.T. (2022), "Local infrastructure, rural households' resilience capacity and poverty: evidence from panel data for Southeast Asia", Journal of Economics and Development, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JED-10-2022-0199

## Publisher

:

Emerald Publishing Limited

Copyright © 2022, Tim Hartwig and Trung Thanh Nguyen

## License

Published in the Journal of Economics and Development. Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) license. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this license may be seen at http://creativecommons.org/licences/by/4.0/ legalcode

## 1. Introduction

Infrastructure development is a topic of paramount interest regarding national, regional and local economic growth. Infrastructure plays a vital role in enabling economic activities in the first place (Esfahani and Ramírez, 2003; Calderón and Servén, 2004; Égert et al., 2009; Chatterjee and Turnovsky, 2012; Daido and Tabata, 2013). In the course of development, the concern about infrastructure and its role in poverty reduction has also been raised (Ali and Pernia, 2003). In developing countries, infrastructure investments are expanding dramatically to boost economic growth and reduce poverty. For instance, Southeast Asia is one of the regions in the world advancing the fastest, both in terms of infrastructure and economic development, and countries in this region, such as Thailand and Vietnam, have experienced rapid economic growth in the last decades (World Bank, 2021a).

However, growth in national products per capita does not necessarily imply widespread economic integration across a country or region (Calderón and Servén, 2010), especially in emerging economies. For example, the metropolitan areas in Thailand saw steady wealth creation and poverty alleviation, while the rural areas profited much less from this development (FAO, 1998). About 20 years ago, 75% of the Vietnamese and 69% of the Thai population lived in rural areas. Today, roughly 65% of the Vietnamese population and 49% of the Thai population live in rural areas. As a result, a clear trend of urbanization has been observed in these countries since the share of people living in rural areas is declining significantly (World Bank, 2021b). On the other hand, this trend increases the inequality and gap between rural and urban areas in terms of income, consumption and living conditions (Do and Park, 2019; Hoang, 2020; Obermann et al., 2020).

This situation could even be worsened through exposure to various exogenous shocks with severe economic impacts, especially in Southeast Asia (Nguyen and Nguyen, 2020; Nguyen et al., 2020a, 2020b). This region is severely affected by natural disasters. US$91 bn equivalent losses are estimated alone due to natural disasters, like storms, floods or droughts, between 2004 and 2014 (ADB, 2021). Economic shocks, like the Asian financial crisis or the world financial crisis, drove millions who had been lifted out of poverty years ago back into it (Habib et al., 2010). The extraordinary events, such as the Covid-19 pandemic, have forced millions of people in the Greater Mekong Region, who had migrated from rural areas to the cities for working opportunities, to return to their home villages (Waibel et al., 2020). As a result, the demand for food and other necessities in rural areas has increased, while remittances from migrant workers have been decreasing. Since monetary means are scarce, especially in rural areas, finding a good way of steering the available resources to help build resilience against exogenous shocks while at the same time not hampering but rather fueling long-term economic development is of critical importance. The available capital stock should be employed so that money is spent in the most efficient way possible. One promising area of investment, therefore, is infrastructure development. By building the backbone of every advanced economy, infrastructure plays a crucial role in fighting poverty and lifting the standard of living of people and households. Evidence suggests that roads and extended irrigation mainly contribute to poverty reduction and economic development (Ali and Pernia, 2003). Nevertheless, it is crucial to highlight how infrastructure development can help improve a household's resilience capacity to fight against these exogenous shocks. While studies on investments in infrastructure contribute and economic development are rich, evidence on the association between infrastructure, household resilience capacity and poverty is still nearly unexplored. Examining this association is essential. Since high volatility in income and consumption can put significant stress on the well-being of the affected population, they may seek less profitable and less volatile income sources, which leads to smoothened consumption and a lower standard of living (Klasen and Waibel, 2015). Furthermore, monitoring the present state of poverty is an essential practice to tailor poverty reduction policies. Against this background, this study aims at answering two research questions: RQ1. How infrastructure has an impact on a household's resilience capacity against shocks and RQ2. How infrastructure and a household's resilience capacity affect poverty and vulnerability to poverty. The remaining of this paper is organized as follows. Section 2 describes the conceptual framework and reviews the related literature. Section 3 introduces the study sites and describes the data. Section 4 explains our research methods. Section 5 presents key results and discusses the findings. Finally, Section 6 concludes with policy implications. ## 2. Conceptual framework and literature review ### 2.1 Conceptual framework This study relies on the sustainable livelihoods framework proposed by Ashley and Carney (1999) and Ali and Pernia (2003) to link local infrastructure development to household livelihood strategies and poverty. In this regard, under the context of shocks, infrastructure development can assist rural households in improving their resilience capacity to deal with shocks, sustain consumption and prevent them from falling into poverty. Neuman (2006) describes infrastructure as a physical network that enables a substance or information to flow to a place of human activity. It connects producers, service providers and users and uses standardized technologies, pricing and controls. While infrastructure undoubtedly builds the backbone of every advanced economy, the causal links between infrastructure and economic development have been discussed until today (Välilä, 2020). Although numerous studies have analyzed the links between infrastructure and economic growth and development, the evidence on this relationship remains mixed. This problem can be attributed to multiple obstacles that emerge when analyzing infrastructure and its impact on economic outcomes (Du et al., 2022). New infrastructure may seem to have little to do with rural development since it requires heavy investment and a well-developed infrastructure foundation even though there are conclusions that could also be driven by rural development. First, quality economic growth as the desired outcome is highly relevant for rural development since growth in terms of gross domestic product (GDP) per capita does not always imply interregional growth and economic integration but can sometimes be attributed to, for example, growth in urban areas but not the rural areas (OECD, 2022). Second, technological innovation and productivity growth as channels for economic development are promising tools when trying to alleviate poverty (World Bank, 2020). Therefore, incorporating these dimensions when conceptualizing the link between infrastructure and rural development may show important interlinkages incorporated in this relationship. Infrastructure investments are thereby subdivided into economic and social infrastructure, reducing poverty through productivity gains and social welfare (Fagbemi et al., 2022). Public investments in economic and social infrastructure, especially in an economy characterized by imperfect markets like they often occur in developing and transition countries, should lift national income and employment and improve social welfare. The results imply that public infrastructure investments help fight poverty and, in turn, lower poverty enhances performance in the public sector, leading to complementary effects. ### 2.2 Literature review Empirical evidence on infrastructure and its effect on rural development is somewhat mixed. Infrastructure investments can help fuel economic growth and regional economic development and reduce poverty (Esfahani and Ramírez, 2003; Chatterjee and Turnovsky, 2012; Daido and Tabata, 2013). However, economic growth does not always imply economic integration across all regions inside an economy. For example, sub-Saharan Africa saw steady GDP per capita growth in recent decades. However, the share of the population employed in agriculture, manufacturing and services remained nearly the same as before the economic expansion. The cost of doing business is among the highest in the world, implying the regional lack of economic transformation (Ajakaiye and Ncube, 2010). Higher investments in infrastructure development, such as transportation networks or nonfarm employment opportunities, and more government oversight in the construction of infrastructure projects may present a viable source of economic transformation, which would also benefit the poor and lead to sustainable economic growth, as it has been shown in Southeast Asia (Do et al., 2022). Investments in other types of infrastructure should also be considered besides electrification (Fan et al., 2004). Studies on the linkage between infrastructure and household resilience are relatively scarce. From the literature, the dominant conceptualization of resilience considers it as a capacity with ex ante attributes (Béné et al., 2012), and the vulnerable context, especially in rural areas, influences households' strategies to build up their resilience capacity to prevent, mitigate or cope with risks (Meybeck et al., 2012). These resilience strategies help rural households sustain their welfare in the short term, prevent them from falling into poverty and reduce their vulnerability to poverty in the long term. At this point, it is essential to point out the difference between poverty and vulnerability to poverty since these are interlinked but very different concepts. Absolute poverty is an income below a fixed poverty line, while relative poverty refers to an income below a certain level in an economy (Foster, 1998). While the first can be used to measure poverty on a global scale, the latter is useful when examining poverty, especially in developed nations, since people living in the developed world can be nonpoor by international standards but be counted as disadvantaged by different country standards. On the other hand, vulnerability to poverty is different from that concept. It could be defined as the ex ante risk that a household will, if currently non-poor, fall below the poverty line, or if currently poor, will remain in poverty” (Jalan et al., 2002, p. 4). This study contributes to the filling of the following research gaps. First, rich findings are attained for irrigation-related infrastructure (Adetoro et al., 2022; Biru et al., 2020; Fischer et al., 2022). However, rural infrastructure includes a broader range of facilities such as roads, electricity and information and communication technology (ICT). We enrich the literature by using more comprehensive indicators of local infrastructure. Second, we offer the first effort to consider the impacts of infrastructure on multidimensional poverty based on the multidimensional poverty measure (MPM) established by the World Bank (2022). Besides the poverty indicators using absolute and relative terms, multidimensional poverty provides a complete picture of the effects of infrastructure development on rural households' welfare. Last, even though some studies have documented a statistically significant influence of infrastructure development on vulnerability to poverty (Leichenko and Silva, 2014; Herrera et al., 2018), they are mainly case studies within a country. Our study is from two countries' rural areas, thus allowing for a better generalization of the findings. ## 3. Study sites and data description ### 3.1 Study sites This study uses the data from the Thailand Vietnam Socio Economic Panel (TVSEP): poverty dynamics and sustainable development project (www.tvsep.de) funded by the German Research Foundation to provide high-quality data on livelihoods and poverty dynamics in the rural areas of the two emerging economies of Thailand and Vietnam. The data include about 4,400 households from three provinces in Thailand, namely Buri Ram, Ubon Ratchathani and Nakhon Phanom, and three provinces in Vietnam, namely Ha Tinh, Thua Thien Hue and Dak Lak (Figure 1). The sampling is based on the guidelines of the United Nations Department of Economic and Social Affairs (Nguyen et al., 2021; Nguyen and Do, 2022). The survey instruments include a household questionnaire and a village questionnaire which are available on the project's webpage. The survey sites in Thailand and Vietnam are economically dominated by agriculture. Necessary infrastructure, like roads, is not well established. Therefore, economic integration is inadequate and is doomed to produce poverty. Furthermore, the Vietnamese regions are regularly subject to natural disasters like floods and storms, which further fuels economic deterioration (Nguyen et al., 2022a, b). The final sample for this study consists of 1,698 households in Thailand and 1,701 households in Vietnam from three survey waves conducted in 2010, 2013 and 2016. This makes a total of 5,094 observations for Thailand and 5,103 for Vietnam. Therefore, the whole data set includes 10,197 observations. In addition to the TVSEP data, we use the rainfall data from the Tropical Rainfall Measuring Mission (TRMM). The resolution of the TRMM rainfall data is highly spatial and temporal and available from 1998 to 2014. Since our household and village data are from 2010, 2013 and 2016 (a three-year gap), we follow the study of Do et al. (2022) to use the three-year lag (t-3) of rainfall as an instrumental variable. ### 3.2 Data description There are no missing entries for the variables of interest, and the dataset is therefore balanced (see Table A1 for the definition and measurement of variables). Panel A1 of Table 1 shows household income and consumption statistics. On average, Vietnamese households have lower living standards than Thai households. Panel A2 of Table 1 shows that Vietnamese and Thai households differ substantially in several aspects. Vietnamese households are more likely to be male-headed than Thai households. Furthermore, household members' average age and health status are higher in Thailand than in Vietnam. The average year of education for household members older than 15 years is also higher in Thailand. The total land area and asset per capita are significantly better in Thailand. Panel B of Table 1 shows the descriptive summary of village characteristics. On the one hand, Thai households are more likely to have a public water supply available in their villages and have more all-time accessible roads instead of dirt roads than Vietnamese households. On the other hand, Vietnamese households are slightly more likely to have a bank office in their villages. Furthermore, Thailand's average distance from the villages to the next market is higher. Other differences in village characteristics are not statistically significant. In sum, with a few exceptions, Thai households have better access to basic infrastructure than Vietnamese households. ### 3.3 Poverty measurement We rely on consumption data to measure the poverty of rural households following Haughton and Khandker (2009), Nguyen et al. (2022a) and Forster (1998). We use several indicators of absolute poverty, relative poverty and multidimensional poverty. An absolute poverty threshold of PPP$ 3.20 per capita daily is applied, as the World Bank (2022) suggested for middle-income countries. A household with expenditure per capita 30% lower than the average per capita is classified as relatively poor. The multidimensional poverty in our study is adjusted from the multidimensional poverty developed by the World Bank (World Bank, 2022). It covers four dimensions, monetary poverty, education poverty, lacking access to basic infrastructure and housing poverty. Each dimension is thereby weighted with 1/4. The cut-off value for poverty is set to 0.25, implying that a household is multidimensionally poor if its parameters add up to 0.25 or higher (see Table A2 for detailed parameters and weights).

Table 2 presents the descriptive summary of multidimensional poverty and poverty indicators. In terms of multidimensional poverty, Thai households have more proper sanitation or access to drinking water than Vietnamese households. However, Vietnamese households have better access to education, at least in quantitative terms. Vietnamese households are less likely to have at least one school-age child who is not enrolled in school or to have at least one adult who has not completed primary education.

## 4. Methodology

We focus on the absorptive capacity of rural households. This resilience capacity denotes the households' ability to prevent potential shocks, mitigate shock impacts and recover quickly from shocks (Béné et al., 2016; Meybeck et al., 2012). Theoretically, consumption smoothing is essential in dealing with shocks. Thus, we assume that a good approximation for the ability to cope with shocks is whether the household had to reduce its consumption due to shocks or not (Nguyen et al., 2022a). In other words, if a household has a resilience capacity to deal with shocks, it does not have to reduce consumption when faced with shocks. Besides, the economic capital of households might play a role in dealing with shocks. For this purpose, two dummy variables are employed to capture households' absorptive capacity, namely (1) reduced consumption due to shocks (if a household had to reduce its consumption to cope with shock = 1; otherwise = 1) and (2) weak economic capacity (if the ratio of household income to poverty threshold [at PPP3.20 per capita per day] is less than one = 1; otherwise = 0). The model of infrastructure and households' resilience capacity can then be expressed as follows: (1)Rit=α1+α2Hit+α3Vit+εit where Rit represents the ability to cope with shocks of household i at time t; Rit includes two dummy variables: reduced consumption due to shocks and weak economic capacity; Hit is the group of household characteristics including gender, age, health status of the household head, whether the head was born in the same (as current) village, mean schooling years of the household's adult members, share of laborers, land area per capita, whether the household has productive machines and motorcycles and whether it is asset poor (belonging to the 20% poorest of asset value per capita); Vit captures the village's infrastructure characteristics which include the number of enterprises in village that provide off-farm employment opportunities, whether the village has made roads instead of dirt roads, the share of households with electricity at home, share of households with a phone line at home, share of households with cable Internet at home, whether the village has public water supply, whether there is a bank office/branch in the village and the distance from the village to the closest market. These households and village variables have been found to influence on rural households' livelihood strategies significantly (Do et al., 2022; Nguyen et al., 2020a, 2022c); and εit is the error term. There might be a concern with using a fixed-effects linear probability model (FE-LPM) for a binary dependent variable. However, the FE-LPM has been found to yield better results in rare events where the number of observations with values of one is smaller than 25% (Timoneda, 2021). Since our descriptive statistics have shown that only 16% of all households had reduced consumption due to shocks in the case of Thailand, the FE-LPM might outperform logistic regressions. As a robustness check, we also run equation (1) with a random-effects Probit model (RE-Probit. We check multicollinearity with the variance inflation factor (VIF) values. The results of VIF values show no significant signs of this problem (see column (1) of Table A2 for the detailed results of VIF values). We cluster our estimations at the village level. ### 4.2 Examining the impacts of resilience capacity on consumption and poverty In this step, we estimate the impact of resilience capacity on consumption, poverty and vulnerability to poverty. The dependent variables include daily household expenditure per capita, household poverty (in absolute poverty, relative poverty and multidimensional poverty) and household vulnerability to poverty. While using the first two groups of dependent variables is clear, we need to construct the indicators of household vulnerability to poverty. Since vulnerability to poverty is an ex ante measure, a multiperiod measure is added, i.e. whether a household, which is currently poor, has been poor in the previous period. Therefore, three dummy variables are generated, namely (1) chronic absolute poverty, (2) chronic relative poverty and (3) chronic multidimensional poverty (yes = 1; otherwise = 0). The fixed-effects model to evaluate the impacts of households' resilience capacity can be specified as follows: (2)Yit=β1+β2Rit+β3Hit+β4Vit+ϵit where Yit denotes three groups of households' welfare, namely (1) household daily expenditure per capita, (2) household poverty (in absolute poverty, relative poverty and multidimensional poverty) and (3) household vulnerability to poverty (chronic absolute poverty, chronic relative poverty and chronic multidimensional poverty); Rit is a vector of the two variables of resilience capacity; Hit and Vit represent household and village characteristics and ϵit is the error term. Since Rit is apparently endogenous, we address this problem by using an instrumental approach (IV). We use a fixed-effects estimation with the IV to estimate equation (2). To instrument Rit, we use the three-year lag with extreme precipitation from the TRMM data. We conduct two quality tests to check whether the IV is appropriate: weak identification (Stock and Yogo, 2005) and underidentification (Cragg and Donald, 1993). The results of these tests presented in the postestimation part of Table 4, and Tables A4–A6 show that our IV estimations do not suffer the problem of underidentification and weak-identification. We further check for the multicollinearity problem of independent variables using the VIF values. The results of VIF values do not imply this problem in our model (see columns (2) and (3) of Table A3). We cluster our estimations at the village level. ## 5. Results and discussion ### 5.1 Association between infrastructure and resilience capacity Table 3 shows the estimation results from FE-LPM and RE-Probit models. The FE-LPM estimation results show statistically significant evidence for the number of enterprises in villages if the village has made roads instead of dirt roads, the share of households with a phone line at home and the share of households having cable Internet at home. All significant coefficients have a negative sign, implying they have a negative association with consumption reduction due to shocks. These results align with previous studies on the role of transportation and ICT infrastructure (Do et al., 2022; Nguyen et al., 2022c). The influence of electricity is, however, less pronounced. This could be because almost all villages in these countries already have access to electricity. Access to public water supply also negatively correlates with consumption reduction in RE-Probit estimations. Our results imply that infrastructure development should focus on transportation, ICT and living facilities. Regarding household variables, household head age, whether the head is healthy, mean schooling years of adult members, owning more motorcycles and the land area per capita have a significant negative association with consumption reduction due to shocks. This implies that improving these characteristics of households helps improve their capacity to cope with shocks. The important role of education in rural households is consistent with the study of Ninh (2021). Large household size and asset-poor households appear to have a significant and positive correlation between consumption reduction due to shock. This denotes that these households should be supported to improve their resilience capacity. ### 5.2 Impacts of resilience capacity on consumption and poverty Infrastructure development increases the ability to cope with shocks by improving their resilience capacity. The increased ability to cope with shocks then reduces the vulnerability to poverty. In this way, the impact of infrastructure development on households' vulnerability to poverty is estimated. Panel A of Table 4 depicts the impact of reduced consumption due to shocks on daily expenditure per capita (in logarithm), absolute expenditure poverty, relative expenditure poverty and multidimensional poverty. It appears that having weak resilience capacity in the form of reduced consumption to cope with shocks negatively affects household expenditure per capita and positively affects poverty indicators. Similarly, the results of the impact of a weak economic capacity in panel B of Table 4 remain consistent, implying that households' absorptive capacity is essential to prevent them from reducing their consumption and falling into poverty in different measures. The importance of improving resilience capacity is consistent with the previous studies (Arslan et al., 2018; Khandker, 2012). The impacts on chronic poverty are reported in Table A6 and are in line with those from Ansah et al. (2021), DeLoach and Smith-Lin (2018) and Yilma et al. (2014) regarding the role of the economic capacity of rural households in coping with shocks. Therefore, supportive policies on improving the household's absorptive capacity are strongly recommended. ## 6. Summary and policy implications In this study, we examined the correlation of infrastructure with a household's resilience capacity (absorptive capacity) against shocks and the impacts of this resilience capacity on household consumption and poverty. We used a sample of 1,698 households in Thailand and 1,701 households in Vietnam, the two emerging economies in Southeast Asia. Our study points out some significant findings. First, the infrastructure helps improve a household's resilience capacity. Particularly, access to cable Internet is estimated to have the largest positive influence on coping with shocks, both in terms of weak economic ability and whether a household had to reduce consumption due to shocks. This relationship, however, is likely to be magnified for the already better-off household. Besides, access to cable Internet, made roads and phone lines significantly increased a household's ability to cope with shocks. Since especially made roads have also been shown to decrease poverty and fuel long-term economic development, increasing efforts to enhance the road network could improve the household's ability to cope with shocks in the worse-off remote areas. Therefore, infrastructure development projects should pay more attention to transportation and ICT facilities, especially in those countries with widespread access to electricity. Second, better education of adult members improves the household's resilience capacity, while a larger household size and asset-poor decrease the capacity of rural households to deal with shocks. Therefore, it is recommended that rural education be promoted, and focuses should be placed on poor households vulnerable to shocks and unable to cope with shocks. Third, the impacts of weak economic ability on chronic poverty are positive and significant in the case of absolute and multidimensional poverty. Hence, infrastructure development is recommended to provide job opportunities, generate household income and improve economic capita. Our study still has a number of limitations. First, we employed fixed-effects estimations with an instrumental variable to account for unobserved heterogeneity of household characteristics and endogeneity of the household's resilience capacity. However, the use of FE-LPMs limits the interpretation of our results, for example, the magnitude of the impacts of better resilience capacity on household consumption and poverty. Second, we used individual indicators to represent the absorptive capacity of a household's resilience, while there are several distinct capacities. Therefore, future studies should consider both absorptive capacity and other capacities, such as adaptive or transformative capacity. ## Figures ### Figure 1 Survey sites of the Thailand Vietnam Socio Economic Panel (TVSEP) ## Table 1 Descriptive statistics of surveyed households and villages Whole sample (n = 10,197)201020132016 Thailand (n = 1,698)Vietnam (n = 1,701)Thailand (n = 1,698)Vietnam (n = 1,701)Thailand (n = 1,698)Vietnam (n = 1,701) A. Household variables A1. Consumption and income Reduced consumption due to shocks (yes = 1)0.310.160.54***,b0.130.45***,b0.180.39***,b (0.46)(0.37)(0.50)(0.33)(0.50)(0.39)(0.49) Daily consumption per capita (PPP)5.034.713.41***,a5.704.05***,a7.205.10***,a
(4.28)(3.90)(2.36)(4.65)(3.15)(5.63)(4.13)
Daily income per capita (PPP$)6.776.533.90***,a7.484.84***,a10.447.4***,a (20.34)(19.45)(5.52)(23.27)(6.46)(37.38)(8.30) A2. Household characteristics Gender of household head (male = 1)0.770.740.85***,b0.710.82***,b0.670.80***,b (0.42)(0.44)(0.35)(0.45)(0.38)(0.47)(0.40) Age of household head (years)56.0757.2650.23***,a59.3553.21***,a61.0855.29***,a (12.91)(12.36)(12.85)(12.16)(12.78)(11.63)(12.47) Health status of household head (healthy = 1)0.810.850.74***,b0.840.66***,b0.920.83***,b (0.40)(0.35)(0.44)(0.36)(0.47)(0.28)(0.37) Local household (yes = 1)0.630.600.63*,b0.610.64**,b0.630.64b (0.48)(0.49)(0.48)(0.49)(0.48)(0.48)(0.48) Mean schooling years of adult members (years)5.696.266.14a5.835.33***,a5.405.18**,a (2.66)(2.14)(2.78)(2.39)(2.84)(2.67)(2.90) Household size (number of persons)4.014.144.34***,a3.984.06a3.753.82a (1.70)(1.73)(1.72)(1.70)(1.72)(1.64)(1.64) Share of laborers (%)74.9370.7570.62a72.0274.11***,a83.1078.97***,a (23.03)(22.36)(23.14)(22.74)(23.13)(22.02)(21.96) Household land area per capita (hectares)0.650.980.25***,a1.130.31***,a0.860.35***,a (1.04)(1.17)(0.71)(1.42)(0.65)(1.04)(0.55) Agricultural machines (yes = 1)0.570.460.64***,b0.450.65***,b0.550.69***,b (0.49)(0.50)(0.48)(0.50)(0.48)(0.50)(0.46) Number of motorcycles1.341.310.98***,a1.471.25***,a1.511.52a (0.99)(0.87)(0.85)(1.00)(1.01)(1.00)(1.07) Asset per capita (PPP$)1531.651675.06591.65***,a2399.09871.64***,a2705.70950.6***,a
(3341.52)(3383.15)(760.68)(5048.09)(1359.12)(4648.95)(1515.72)
B. Village variables
Number of enterprises0.460.120.13a0.460.99***,a0.280.79***,a
(1.62)(0.59)(0.56)(1.98)(2.51)(0.88)(1.88)
Having made roads instead of dirt roads (yes = 1)0.840.890.67***,b0.970.64***,b0.960.89***,b
(0.37)(0.32)(0.47)(0.18)(0.48)(0.20)(0.31)
Share of households with electricity at home (%)98.5098.7298.30**,a98.6597.53***,a99.0698.75a
(7.30)(4.14)(7.35)(6.88)(10.86)(4.18)(8.02)
Share of households with a phone line at home (%)82.6037.2879.27***,a98.8988.03***,a99.0693.02***,a
(31.24)(46.34)(19.74)(5.84)(18.75)(4.77)(13.26)
Share of households with cable Internet at home (%)4.451.911.92a3.514.91***,a4.0910.37***,a
(9.50)(5.00)(6.02)(10.46)(7.86)(7.19)(14.44)
Having access to public water supply (yes = 1)0.590.950.30***,b0.920.20***,b0.950.21***,b
(0.49)(0.22)(0.46)(0.28)(0.40)(0.22)(0.41)
Having a bank agency in village (yes = 1)0.060.000.07***,b0.090.06***,b0.060.05b
(0.23)(0.00)(0.26)(0.29)(0.23)(0.25)(0.23)
Distance to next market (kilometers)5.978.942.98***,a8.962.98***,a8.963.00***,a
(6.93)(7.86)(4.40)(7.85)(4.40)(7.33)(4.41)

Note(s): Standard deviations in parentheses; a: two-sample t-test; b: nonparametric rank-sum test; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1

## Table 2

Descriptive statistics on multidimensional poverty and poverty indicators

Whole sample (n = 10,197)201020132016
Thailand (n = 1,698)Vietnam (n = 1,701)Thailand (n = 1,698)Vietnam (n = 1,701)Thailand (n = 1,698)Vietnam (n = 1,701)
MPI indicators
No child education (yes = 1)0.060.060.05b0.110.08***,b0.020.03b
(0.24)(0.24)(0.22)(0.31)(0.27)(0.15)(0.16)
No adult education (yes = 1)0.020.040.02***,b0.030.01***,b0.020.01***,b
(0.15)(0.20)(0.14)(0.18)(0.11)(0.13)(0.09)
No safe drinking water (yes = 1)0.380.160.67***,b0.120.69***,b0.040.60***,b
(0.49)(0.37)(0.47)(0.33)(0.46)(0.20)(0.49)
No improved sanitation (yes = 1)0.300.030.70***,b0.020.59***,b0.010.45***,b
(0.46)(0.17)(0.46)(0.15)(0.49)(0.09)(0.50)
No access to electricity for lighting (yes = 1)0.020.020.01**,b0.040.02**,b0.010.02***,b
(0.14)(0.15)(0.11)(0.18)(0.15)(0.10)(0.15)
No appropriate housing condition (yes = 1)0.160.110.30***,b0.200.19b0.050.11***,b
(0.37)(0.31)(0.46)(0.40)(0.39)(0.23)(0.31)
No appropriate nutrition for children (yes = 1)0.120.140.18***,b0.110.15***,b0.070.10***,b
(0.33)(0.34)(0.38)(0.31)(0.36)(0.25)(0.30)
Poverty indicators
Absolute consumption poverty (yes = 1)0.370.390.55***,b0.280.46***,b0.170.35***,b
(0.48)(0.49)(0.50)(0.45)(0.50)(0.37)(0.48)
Relative consumption poverty (yes = 1)0.390.410.35***,b0.410.39b0.420.39*,b
(0.49)(0.49)(0.48)(0.49)(0.49)(0.49)(0.49)
Multidimensional poverty (yes = 1)0.400.410.61***,b0.310.53***,b0.170.38***,b
(0.49)(0.49)(0.49)(0.46)(0.50)(0.38)(0.49)

Note(s): Standard deviations in parentheses; a: two-sample t-test; b: nonparametric rank-sum test; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1; MPI: Multidimensional Poverty Index

## Table 3

Association between infrastructure development and resilience capacity

Reduced consumption due to shocksWeak economic capacity
FE-LPMRE-ProbitFE-LPMRE-Probit
Male head−0.0170.0070.043−0.000
(0.027)(0.040)(0.030)(0.037)
Age of head−0.002*−0.012***−0.007***−0.007***
(0.001)(0.001)(0.001)(0.001)
Healthy head−0.066***−0.316***−0.053***−0.250***
(0.016)(0.040)(0.015)(0.040)
Head born in the village0.0100.030−0.0070.098***
(0.028)(0.037)(0.031)(0.034)
Mean schooling years of adult members0.002−0.014**−0.005*−0.059***
(0.003)(0.006)(0.003)(0.007)
Household size0.016***0.068***0.047***0.137***
(0.005)(0.011)(0.006)(0.012)
Share of laborers−0.0000.002***−0.002***−0.006***
(0.000)(0.001)(0.000)(0.001)
Land area per capita−0.014*−0.060***0.003−0.082***
(0.007)(0.023)(0.007)(0.021)
Having productive machines0.030**0.112***−0.024*−0.130***
(0.014)(0.039)(0.015)(0.033)
Number of motorcycles−0.006−0.103***−0.062***−0.326***
(0.007)(0.020)(0.008)(0.020)
Asset poor0.0110.098**0.051***0.381***
(0.017)(0.044)(0.017)(0.041)
Number of enterprises−0.005*−0.0070.002−0.007
(0.003)(0.010)(0.004)(0.011)
Having made roads instead of dirt roads−0.042*−0.208***−0.074***−0.237***
(0.023)(0.052)(0.023)(0.055)
Share of households with electricity at home (%)0.000−0.0030.000−0.001
(0.001)(0.002)(0.001)(0.002)
Share of households with a phone line at home (%)−0.000**−0.001**−0.000−0.001***
(0.000)(0.001)(0.000)(0.001)
Share of households with cable Internet at home (%)−0.002***−0.008***−0.002**−0.009***
(0.001)(0.002)(0.001)(0.002)
Having access to public water supply0.012−0.508***0.022−0.121***
(0.023)(0.046)(0.025)(0.043)
Having a bank agency in village−0.032−0.0500.0240.053
(0.029)(0.071)(0.039)(0.086)
Distance to the closest market0.001−0.011***0.0020.002
(0.003)(0.004)(0.002)(0.003)
Constant0.432***1.010***0.959***1.558***
(0.123)(0.275)(0.104)(0.238)
Number of observations10,19710,19710,19710,197
Prob > F0.000 0.000
Prob > χ2 0.000 0.000

Note(s): Robust standard errors clustered at village level; : dummy; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1

## Table 4

Impact of resilience capacity on household consumption and poverty

Expenditure per capita (ln)Absolute expenditure povertyRelative expenditure povertyMultidimensional poverty
A. Estimations on reduced consumption against shocks
Reduced consumption due to shocks−0.942***0.969***0.478***0.910***
(0.278)(0.253)(0.179)(0.256)
Household variablesYesYesYesYes
Village variablesYesYesYesYes
Constant1.365***−0.019−0.398***0.070
(0.191)(0.166)(0.133)(0.171)
Number of observations10,19710,19710,19710,197
Wald χ2(20)1301.342485.681526.866520.477
Prob > χ20.0000.0000.0000.000
Underidentification0.0000.0000.0000.000
Weak identification23.31823.31823.31823.318
B. Estimations on weak economic capacity
Weak economic capacity−0.697***0.717***0.353***0.673***
(0.150)(0.141)(0.127)(0.143)
Household variablesYesYesYesYes
Village variablesYesYesYesYes
Constant1.627***−0.289*−0.531***−0.183
(0.183)(0.167)(0.158)(0.162)
Number of observations10,19710,19710,19710,197
Wald χ2(20)2434.245800.012606.026907.550
Prob > χ20.0000.0000.0000.000
Underidentification0.0000.0000.0000.000
Weak identification42.94242.94242.94242.942

Note(s): Robust standard errors clustered at village level; : dummy; ln: natural logarithm; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1; the underidentification test is an LM test based on the rk LM statistics. The null hypothesis of this LM test is that the model is underidentified. The reported weak identification test is the Kleibergen-Paap rk Wald F-statistic. Full results are presented in Tables A4–A5

## Table A1

Definition and measurement of household and village characteristics

VariablesMeasurementDefinition
A. Consumption and income variables
Reduced consumption due to shocksDummyIf the household had to reduce consumption due to shocks in the current year = 1; otherwise = 0
Daily consumption per capitaPPP$(adjusted to 2005 prices)The daily consumption per capita of household Daily income per capitaPPP$ (adjusted to 2005 prices)The daily income per capita of household
B. Multidimensional poverty indicators
No child educationDummyAt least one school-age child up to the age of Grade 8 is not enrolled in school = 1; otherwise = 0
No adult educationDummyNo adult in the household (age of Grade 9 or above) has completed primary education = 1; otherwise = 0
Unsafe drinking waterDummyThe household has drinking water from well, river, lake and pond = 1; otherwise = 0
No improved sanitationDummyThe household has no flush toilets = 1; otherwise = 0
No access to electricityDummyThe household has no access to electricity for lighting = 1; otherwise = 0
No appropriate housing conditionDummyThe household has dwelling size per capita less than 10 m2 = 1; otherwise = 0
No appropriate nutrition for childrenDummyThe household has malnourished child = 1; otherwise = 0
C. Household characteristics
Gender of household headDummyGender of the household head. Male household head = 1; otherwise = 0
Age of household headYearsAge of the household head
Health status of household headDummyIf household head is healthy or “can manage” = 1; otherwise = 0
Local householdDummyIf the household head was born in the same as the current village = 1; otherwise = 0
Mean schooling years of adult membersYearsAverage years of schooling of adult members in the household
Household sizeQuantityNumber of members in the household
Share of laborersPercentageShare of members in working ages (from 15 to 64 years old) in the household
Household land area per capitaHectaresLand area per capita of the household
Agricultural machinesDummyIf the household has agricultural machines including four-wheel tractors, engine sprayers, pumps, tanks, rice mill and threshing machines = 1; otherwise = 0
Number of motorcyclesQuantityNumber of motorcycles that the household owns
Asset per capitaPPP$(adjusted to 2005 prices)Total accumulated asset value per capita of the household D. Village characteristics Number of enterprisesQuantityNumber of enterprises with more than nine employees in the village Having made roads instead of dirt roadsDummyIf the main roads in the village are made roads (instead of dirt roads) = 1; otherwise = 0 Share of households with electricity at homePercentageThe percentage of households with electricity at home in the village Share of households with a phone line at homePercentageThe percentage of households with a phone line at home in the village Share of households with cable Internet at homePercentageThe percentage of households with cable Internet at home in the village Having access to public water supplyDummyIf the village has public water supply available = 1; otherwise = 0 Having a bank agency in villageDummyIf the village has a bank agency available = 1; otherwise = 0 Distance to next marketKilometersThe distance from the household to the next market if in village ## Table A2 Adjusted multidimensional poverty measure DimensionParameterWeight MonetaryDaily consumption is less than US$3.20 PPP per capita1/4
EducationAt least one school-age child up to the age of grade 8 is not enrolled in school1/8
No adult in the household (age of grade 9 or above) has completed primary education1/8
Access to basic infrastructureThe household lacks access to safe sources for drinking water1/12
The household lacks access to flush toilets1/12
The household has no access to electricity for lighting1/12
HousingThe household has a dwelling size of less than 10 m2 per capita1/8
The household has a malnourished child1/8

Source(s): Based on World Bank (2022)

## Table A3

Variance inflator factor (VIF) of independent variables

Correlation of infrastructure and household's resilience capacityImpacts of household resilience capacity on poverty
Reduced consumption due to shocksWeak economic capacity
(1)(2)(3)
Reduced consumption due to shocks 1.12
Weak economic capacity 1.21
Male head1.081.081.08
Age of head1.181.191.18
Healthy head1.111.111.11
Head born in the village1.091.091.09
Mean schooling years of adult members1.111.111.12
Household size1.401.411.43
Share of laborers1.231.231.24
Land area per capita1.161.161.17
Having productive machines1.201.201.20
Number of motorcycles1.371.381.41
Asset poor1.251.251.27
Number of enterprises1.051.051.05
Having made roads instead of dirt roads1.221.221.22
Share of households with electricity at home1.021.021.02
Share of households with a phone line at home1.081.081.08
Share of households with cable Internet at home1.101.101.10
Having access to public water supply1.311.341.31
Having access to a bank agency in village1.021.021.02
Distance to the closest market1.181.191.18
Mean VIF1.171.171.18

## Table A4

The influence of reduced consumption due to shocks on household consumption and poverty

Expenditure per capita (ln)Absolute expenditure povertyRelative expenditure povertyMultidimensional poverty
Reduced consumption due to shocks−0.942***0.969***0.478***0.910***
(0.278)(0.253)(0.179)(0.256)
Male head−0.0210.014−0.0030.023
(0.037)(0.035)(0.030)(0.034)
Age of head0.0020.0010.006***−0.001
(0.001)(0.001)(0.001)(0.001)
Healthy head−0.0340.0340.047**0.010
(0.027)(0.025)(0.019)(0.025)
Head born in the village0.009−0.0080.025−0.003
(0.039)(0.034)(0.029)(0.033)
Mean schooling years of adult members−0.0050.002−0.006**−0.000
(0.004)(0.004)(0.003)(0.004)
Household size−0.128***0.073***0.072***0.083***
(0.008)(0.008)(0.006)(0.008)
Share of laborers0.003***−0.002***−0.001**−0.001***
(0.000)(0.000)(0.000)(0.000)
Land area per capita−0.0010.012−0.0030.012
(0.011)(0.009)(0.007)(0.008)
Having productive machines0.107***−0.055***−0.028*−0.059***
(0.021)(0.020)(0.015)(0.019)
Number of motorcycles0.119***−0.072***−0.043***−0.077***
(0.011)(0.009)(0.007)(0.009)
Asset poor−0.128***0.089***0.139***0.072***
(0.022)(0.023)(0.018)(0.022)
Number of enterprises−0.0010.0050.005*0.003
(0.004)(0.004)(0.003)(0.004)
Having made roads instead of dirt roads−0.0080.0370.048**0.034
(0.031)(0.030)(0.020)(0.029)
Share of households with electricity at home (%)0.000−0.0000.000−0.000
(0.001)(0.001)(0.001)(0.001)
Share of households with a phone line at home (%)0.001***−0.001***0.000**−0.001**
(0.000)(0.000)(0.000)(0.000)
Share of households with cable Internet at home (%)0.0010.0010.002***0.001
(0.001)(0.001)(0.001)(0.001)
Having access to public water supply0.0120.002−0.0210.001
(0.028)(0.025)(0.020)(0.026)
Having a bank agency in village−0.0200.015−0.0020.011
(0.051)(0.042)(0.032)(0.039)
Distance to the closest market0.003−0.0000.0020.001
(0.004)(0.003)(0.002)(0.003)
Constant1.365***−0.019−0.398***0.070
(0.191)(0.166)(0.133)(0.171)
Number of observations10,19710,19710,19710,197
Wald χ2(20)1301.342485.681526.866520.477
Prob > χ20.0000.0000.0000.000
Under-identification0.0000.0000.0000.000
Weak-identification23.31823.31823.31823.318

Note(s): Robust standard errors clustered at village level; : dummy; ln: natural logarithm; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1; the underidentification test is an LM (Lagrange multiplier) test based on the rk LM statistics. The null hypothesis of this LM test is that the model is underidentified. The reported weak identification test is the Kleibergen-Paap rk Wald F-statistic

## Table A5

The influence of weak economic capacity on household consumption and poverty

Expenditure per capita (ln)Absolute expenditure povertyRelative expenditure povertyMultidimensional poverty
Weak economic capacity−0.697***0.717***0.353***0.673***
(0.150)(0.141)(0.127)(0.143)
Male head0.026−0.033−0.027−0.022
(0.033)(0.030)(0.029)(0.029)
Age of head−0.0010.004**0.008***0.002
(0.002)(0.001)(0.001)(0.001)
Healthy head−0.0080.0070.034**−0.015
(0.019)(0.018)(0.016)(0.017)
Head born in the village−0.0060.0070.0320.012
(0.032)(0.027)(0.026)(0.028)
Mean schooling years of adult members−0.010***0.007***−0.0030.005*
(0.003)(0.003)(0.002)(0.003)
Household size−0.110***0.054***0.063***0.066***
(0.009)(0.009)(0.008)(0.009)
Share of laborers0.001***−0.0000.000−0.000
(0.000)(0.000)(0.000)(0.000)
Land area per capita0.014−0.004−0.010−0.002
(0.009)(0.006)(0.006)(0.006)
Having productive machines0.062***−0.008−0.006−0.015
(0.015)(0.015)(0.013)(0.014)
Number of motorcycles0.082***−0.033***−0.024**−0.040***
(0.013)(0.011)(0.010)(0.011)
Asset poor−0.102***0.062***0.126***0.047**
(0.018)(0.020)(0.018)(0.019)
Number of enterprises0.005−0.0010.002−0.003
(0.004)(0.003)(0.003)(0.003)
Having made roads instead of dirt roads−0.0210.049**0.055***0.046**
(0.026)(0.023)(0.020)(0.023)
Share of households with electricity at home (%)−0.000−0.0000.0000.000
(0.001)(0.001)(0.001)(0.001)
Share of households with a phone line at home (%)0.002***−0.001***0.000*−0.001***
(0.000)(0.000)(0.000)(0.000)
Share of households with cable Internet at home (%)0.002***−0.0000.002**−0.000
(0.001)(0.001)(0.001)(0.001)
Having access to public water supply0.016−0.002−0.023−0.003
(0.023)(0.021)(0.019)(0.021)
Having a bank agency in village0.026−0.033−0.026−0.034
(0.031)(0.027)(0.026)(0.026)
Distance to the closest market0.003−0.0000.0020.001
(0.003)(0.002)(0.002)(0.002)
Constant1.627***−0.289*−0.531***−0.183
(0.183)(0.167)(0.158)(0.162)
Number of observations10,19710,19710,19710,197
Wald χ2(20)2434.245800.012606.026907.550
Prob > χ20.0000.0000.0000.000
Underidentification0.0000.0000.0000.000
Weak identification42.94242.94242.94242.942

Note(s): Robust standard errors clustered at village level; : dummy; ln: natural logarithm; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1; the underidentification test is an LM test based on the rk LM statistics. The null hypothesis of this LM test is that the model is underidentified. The reported weak identification test is the Kleibergen-Paap rk Wald F-statistic

## Table A6

The influence of the ability to cope with shocks on the vulnerability to poverty

Chronic absolute povertyChronic relative povertyChronic multidimensional povertyChronic absolute povertyChronic relative povertyChronic multidimensional poverty
Reduced consumption due to shocks1.370−0.4711.570
(1.018)(0.495)(1.163)
Weak economic capacity 0.362***−0.1250.418***
(0.101)(0.094)(0.113)
Male head0.091−0.0420.114−0.001−0.0110.008
(0.082)(0.043)(0.095)(0.026)(0.027)(0.028)
Age of head−0.0020.002*−0.0020.0010.0010.001
(0.002)(0.001)(0.003)(0.001)(0.001)(0.001)
Healthy head0.025−0.0060.020−0.0080.006−0.017
(0.054)(0.025)(0.061)(0.016)(0.014)(0.018)
Head born in the village−0.0570.021−0.0530.012−0.0030.026
(0.069)(0.042)(0.079)(0.027)(0.030)(0.029)
Mean schooling years of adult members0.011−0.0020.0110.004*0.0000.003
(0.007)(0.003)(0.008)(0.002)(0.002)(0.003)
Household size0.026*0.022***0.031*0.027***0.022***0.032***
(0.015)(0.008)(0.017)(0.007)(0.006)(0.007)
Share of laborers−0.001−0.001*−0.001−0.000−0.001**−0.000
(0.001)(0.000)(0.001)(0.000)(0.000)(0.000)
Land area per capita0.021−0.0010.0250.0030.0060.004
(0.017)(0.009)(0.019)(0.005)(0.006)(0.007)
Having productive machines−0.086*0.011−0.089−0.019−0.012−0.012
(0.051)(0.024)(0.057)(0.014)(0.013)(0.015)
Number of motorcycles−0.053***−0.009−0.056***−0.026***−0.018**−0.027***
(0.017)(0.008)(0.019)(0.008)(0.007)(0.009)
Asset poor0.0170.076***0.0100.042*0.068***0.034
(0.055)(0.029)(0.062)(0.022)(0.020)(0.021)
Number of enterprises0.003−0.0000.0030.0010.0000.001
(0.005)(0.002)(0.005)(0.003)(0.002)(0.003)
Having made roads instead of dirt roads0.0890.0140.0870.059**0.0240.054**
(0.078)(0.038)(0.091)(0.024)(0.020)(0.026)
Share of households with electricity at home (%)−0.001−0.000−0.001−0.000−0.001−0.000
(0.002)(0.001)(0.002)(0.000)(0.000)(0.001)
Share of households with a phone line at home (%)0.002−0.0010.002−0.0000.000−0.000
(0.002)(0.001)(0.002)(0.000)(0.000)(0.000)
Share of households with cable Internet at home (%)0.001−0.0010.002−0.0000.000−0.000
(0.002)(0.001)(0.002)(0.001)(0.000)(0.001)
Having access to public water supply−0.0070.012−0.0270.0000.010−0.018
(0.066)(0.029)(0.079)(0.024)(0.021)(0.030)
Having a bank agency in village−0.055−0.006−0.054−0.034−0.013−0.031
(0.090)(0.031)(0.096)(0.035)(0.032)(0.038)
Distance to the closest market−0.0070.002−0.006−0.004**0.001−0.002
(0.005)(0.003)(0.006)(0.002)(0.001)(0.002)
Constant−0.2190.324−0.2830.0130.244*−0.019
(0.540)(0.252)(0.610)(0.149)(0.139)(0.171)
Number of observations6,7986,7986,7986,7986,7986,798
Wald χ2(20)58.35862.40853.553282.62696.175293.167
Prob > χ20.0000.0000.0000.0000.0000.000
Underidentification0.1510.1510.1510.0000.0000.000
Weak identification2.0832.0832.08331.93131.93131.931

Note(s): Robust standard errors clustered at village level; : dummy; ln: natural logarithm; ∗∗∗ p < 0.01, ∗∗ p < 0.05 and p < 0.1; the underidentification test is an LM test based on the rk LM statistics. The null hypothesis of this LM test is that the model is underidentified. The reported weak-identification test is the Kleibergen-Paap rk Wald F-statistic

Declaration of competing interest: The authors declare that they have no conflict of interest in this research.

Appendix

## References

ADB (2021), “Six ways Southeast Asia strengthened disaster risk management”, The Asian Development Bank (ADB), available at: https://www.adb.org/news/features/six-ways-southeast-asia-strengthened-disaster-risk-management (accessed 06 October 2022).

Adetoro, A.A., Ngidi, M.S.C., Danso-Abbeam, G., Ojo, T.O. and Ogundeji, A.A. (2022), “Impact of irrigation on welfare and vulnerability to poverty in South African farming households”, Scientific African, Vol. 16, e01177.

Ajakaiye, O. and Ncube, M. (2010), “Infrastructure and economic development in Africa: an overview”, Journal of African Economies, Vol. 19 suppl_1, pp. i3-i12.

Ali, I. and Pernia, E.M. (2003), “Infrastructure and poverty reduction - what is the connection?”, The Asian Development Bank (ADB): Economics and Research Department Policy Brief Series, No. 13.

Ansah, I.G.K., Gardebroek, C. and Ihle, R. (2021), “Shock interactions, coping strategy choices and household food security”, Climate and Development, Vol. 13 No. 5, pp. 414-426.

Arslan, A., Cavatassi, R., Alfani, F., Mccarthy, N., Lipper, L. and Kokwe, M. (2018), “Diversification under climate variability as part of a CSA strategy in rural Zambia”, The Journal of Development Studies, Vol. 54 No. 3, pp. 457-480.

Ashley, C. and Carney, D. (1999), Sustainable Livelihoods: Lessons from Early Experience, Department for International Development, Vol. 7 No. 1, London.

Béné, C., Wood, R.G., Newsham, A. and Davies, M. (2012), “Resilience: new utopia or new tyranny? Reflection about the potentials and limits of the concept of resilience in relation to vulnerability reduction programmes”, IDS Working Papers, No. 405, pp. 1-61.

Béné, C., Headey, D., Haddad, L. and von Grebmer, K. (2016), “Is resilience a useful concept in the context of food security and nutrition programmes? Some conceptual and practical considerations”, Food Security, Vol. 8 No. 1, pp. 123-138.

Biru, W.D., Zeller, M. and Loos, T.K. (2020), “The impact of agricultural technologies on poverty and vulnerability of smallholders in Ethiopia: a panel data analysis”, Social Indicators Research, Vol. 147 No. 2, pp. 517-544.

Calderón, C. and Servén, L. (2004), “The effects of infrastructure development on growth and income distribution”, Policy Research Working Paper No. 3400, World Bank, Washington, D.C.

Calderón, C. and Servén, L. (2010), “Infrastructure and economic development in sub-Saharan Africa”, Journal of African Economies, Vol. 19, pp. i13-i87.

Chatterjee, S. and Turnovsky, S.J. (2012), “Infrastructure and inequality”, European Economic Review, Vol. 56 No. 8, pp. 1730-1745.

Cragg, J.G. and Donald, S.G. (1993), “Testing identifiability and specification in instrumental variable models”, Econometric Theory, Vol. 9 No. 2, pp. 222-240.

Daido, K. and Tabata, K. (2013), “Public infrastructure, production organization, and economic development”, Journal of Macroeconomics, Vol. 38, pp. 330-346.

DeLoach, S.B. and Smith-Lin, M. (2018), “The role of savings and credit in coping with idiosyncratic household shocks”, The Journal of Development Studies, Vol. 54 No. 9, pp. 1513-1533.

Do, M.H. and Park, S.C. (2019), “Impacts of Vietnam's new rural development policy on rural households' income: empirical evidence from the Heckman selection model”, International Review of Public Administration, Vol. 24 No. 4, pp. 229-245.

Do, M.H., Nguyen, T.T., Halkos, G. and Grote, U. (2022), “Non-farm employment, natural resource extraction, and poverty: evidence from household data for rural Vietnam”, Environment, Development and Sustainability, pp. 1-38, doi: 10.1007/s10668-022-02391-7.

Du, X., Zhang, H. and Han, Y. (2022), “How does new infrastructure investment affect economic growth quality? Empirical evidence from China”, Sustainability, Vol. 14 No. 6, p. 3511.

Égert, B., Koźluk, T. and Sutherland, D. (2009), “Infrastructure and growth: empirical evidence”, OECD Economics Department Working Papers, No. 685.

Esfahani, H.S. and Ramírez, M.T. (2003), “Institutions, infrastructure, and economic growth”, Journal of Development Economics, Vol. 70 No. 2, pp. 443-477.

Fagbemi, F., Osinubi, T.T. and Adeosun, O.A. (2022), “Enhancing sustainable infrastructure development: a boon to poverty reduction in Nigeria”, World Development Sustainability, Vol. 1, 100006.

Fan, S., Jitsuchon, S. and Methakunnavut, N. (2004), “The importance of public investment for reducing rural poverty in middle-income countries: the case of Thailand”, DSGD Discussion Paper No. 7.

FAO (1998), “Poverty alleviation and food security in Asia: lessons and challenges”, Regional Office for Asia and the Pacific, available at: https://www.fao.org/3/ab981e/ab981e00.htm#Contents (accessed 06 October 2022).

Fischer, C., Aubron, C., Trouvé, A., Sekhar, M. and Ruiz, L. (2022), “Groundwater irrigation reduces overall poverty but increases socioeconomic vulnerability in a semiarid region of southern India”, Scientific Reports, Vol. 12 No. 1, pp. 1-16.

Foster, J. (1998), “Absolute versus relative poverty”, The American Economic Review, Vol. 88, pp. 335-341.

Habib, B., Narayan, A., Olivieri, S. and Sanchez, C. (2010), “The impact of the financial crisis on poverty and income distribution: insights from simulations in selected countries”, Economic Premise No. 7, World Bank, Washington, DC.

Haughton, J. and Khandker, S. (2009), Handbook on Poverty and Inequality, World Bank, Washington, DC.

Herrera, C., Ruben, R. and Dijkstra, G. (2018), “Climate variability and vulnerability to poverty in Nicaragua”, Journal of Environmental Economics and Policy, Vol. 7 No. 3, pp. 324-344.

Hoang, Q.V. (2020), “Determinants of the result of new rural development program in Vietnam”, Journal of Economics and Development, Vol. 22 No. 1, pp. 81-90.

Jalan, J., Chaudhuri, S. and Suryahadi, A. (2002), Assessing Household Vulnerability to Poverty from Cross-Sectional Data: A Methodology and Estimates from Indonesia, Department of Economics, Columbia University, NY.

Khandker, S.R. (2012), “Seasonality of income and poverty in Bangladesh”, Journal of Development Economics, Vol. 97 No. 2, pp. 244-256.

Klasen, S. and Waibel, H. (2015), “Vulnerability to poverty in South-East Asia: drivers, measurement, responses, and policy issues”, World Development, Vol. 71, p. 1.

Leichenko, R. and Silva, J.A. (2014), “Climate change and poverty: vulnerability, impacts, and alleviation strategies”, Wiley Interdisciplinary Reviews: Climate Change, Vol. 5 No. 4, pp. 539-556.

Meybeck, A. Lankoski, J. Redfern, S. Azzu, N. and Gitz, V. (2012), “Building resilience for adaptation to climate change in the agriculture sector”, in Building Resilience for Adaptation to Climate Change in the Agriculture Sector, Proceedings of a Joint FAO/OECD Workshop, The Food and Agriculture Organization of the United Nations (FAO), Rome, 23-24 April 2012.

Neuman, M. (2006), “Infiltrating infrastructures: on the nature of networked infrastructure”, Journal of Urban Technology, Vol. 13 No. 1, pp. 3-31.

Nguyen, T.T. and Do, M.H. (2022), “Female rural–urban migrants and online marketplaces in emerging economies: evidence from Thailand and Vietnam”, Asia and the Pacific Policy Studies, Vol. 9 No. 3, pp. 317-342.

Nguyen, G. and Nguyen, T.T. (2020), “Exposure to weather shocks: a comparison between self-reported record and extreme weather data”, Economic Analysis and Policy, Vol. 65, pp. 117-138.

Nguyen, T.T., Do, M.H. and Rahut, D. (2022b), “Shock, risk attitude and rice farming: evidence from panel data for Thailand”, Environmental Challenges, Vol. 6, 100430.

Nguyen, T.T., Nguyen, T.T. and Grote, U. (2020a), “Multiple shocks and households' choice of coping strategies in rural Cambodia”, Ecological Economics, Vol. 167, 106442.

Nguyen, D.L., Nguyen, T.T. and Grote, U. (2022a), “Shocks, household consumption, and livelihood diversification: a comparative evidence from panel data in rural Thailand and Vietnam”, Economic Change and Restructuring, pp. 1-33.

Nguyen, T.T., Nguyen, T.T. and Grote, U. (2022c), “Internet use, natural resource extraction and poverty reduction in rural Thailand”, Ecological Economics, Vol. 196, 107417.

Nguyen, T.T., Nguyen, T.T., Le, V.H., Managi, S. and Grote, U. (2020b), “Reported weather shocks and rural household welfare: evidence from panel data in Northeast Thailand and Central Vietnam”, Weather and Climate Extremes, Vol. 30, 100286.

Nguyen, T.T., Tran, V.T., Nguyen, T.T. and Grote, U. (2021), “Farming efficiency, cropland rental market and income effect: evidence from panel data for rural Central Vietnam”, European Review of Agricultural Economics, Vol. 48 No. 1, pp. 207-248.

Ninh, L.K. (2021), “Economic role of education in agriculture: evidence from rural Vietnam”, Journal of Economics and Development, Vol. 23 No. 1, pp. 47-58.

Obermann, G., Oanh, N.H. and Ngoc, N.H. (2020), “Gender pay gap in Vietnam: a propensity score matching analysis”, Journal of Economics and Development, Vol. 23 No. 3, pp. 238-253.

OECD (2022), “Thailand's development trajectory: past and future strategies”, available at: https://www.oecd-ilibrary.org/sites/a21090a7-en/index.html?itemId=/content/component/a21090a7-en (accessed 06 October 2022).

Stock, J.H. and Yogo, M. (2005), “Testing for weak instruments in linear IV regression”, in Andrews, D.W.K. and Stock, J.H. (Eds), Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg, Cambridge University Press, Cambridge, pp. 80-108.

Timoneda, J.C. (2021), “Estimating group fixed effects in panel data with a binary dependent variable: how the LPM outperforms logistic regression in rare events data”, Social Science Research, Vol. 93, 102486.

Välilä, T. (2020), “Infrastructure and growth: a survey of macro-econometric research”, Structural Change and Economic Dynamics, Vol. 53, pp. 39-49.

Waibel, H., Grote, U., Min, S., Nguyen, T.T. and Praneetvatakul, S. (2020), “COVID-19 in the Greater Mekong Subregion: how resilient are rural households?”, Food Security, Vol. 12 No. 4, pp. 779-782.

World Bank (2020), “Productivity growth, key driver of poverty reduction, threatened by COVID-19 disruptions”, available at: https://www.worldbank.org/en/news/press-release/2020/07/14/productivity-growth-threatened-by-covid-19-disruptions (accessed 19 May 2022).

World Bank (2021a), “The World Bank open data: GDP growth”, The World Bank Group (WB), available at: https://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG (accessed 06 October 2022).

World Bank (2021b), “The World Bank open data: country indicators”, available at: https://data.worldbank.org/country/ (accessed 06 October 2022).

World Bank (2022), “Multidimensional poverty measure”, available at: https://www.worldbank.org/en/topic/poverty/brief/multidimensional-poverty-measure (accessed 06 October 2022).

Yilma, Z., Mebratie, A., Sparrow, R., Abebaw, D., Dekker, M., Alemu, G. and Bedi, A.S. (2014), “Coping with shocks in rural Ethiopia”, Journal of Development Studies, Vol. 50 No. 7, pp. 1009-1024.

## Acknowledgements

The authors would like to thank the respondents from the surveyed provinces for their kind support and cooperation and Manh Hung Do for his technical support. The authors acknowledge the financial support of the German Research Foundation (DFG - FOR 756/2) for the TVSEP project and appreciate the efforts of their colleagues at the Leibniz University Hannover for data collection and cleaning.

## Corresponding author

Trung Thanh Nguyen can be contacted at: thanh.nguyen@iuw.uni-hannover.de