Long-lasting consequences of war on disability

Abstract

This article investigates the impact of exposure to United States air force bombing during 1965–75 on the disability status of individuals in Vietnam in 2009. Using a combination of national census and US military data and an instrumental variable strategy which exploits the distance to the former North–South border as a quasi-experiment, the article finds a positive and significant impact of bombing exposure on district level disability rates 40 years after the war. The overall effect of bombing on the long-term disability rate among the Vietnamese population is highest among heavily bombed districts. Districts in the top bombing quintile experience a 25% relative increase in the rate of disability attributable to bombing compared with districts in the lowest bombing quintile. Effects are highest on the prevalence of severe disability and among cohorts before the war’s end. A smaller, yet significant, effect is found among cohorts born after the war. The article finds further evidence of indirect channels through which bombing may have impacted on long-term disability including adverse effects on nutritional environment and human capital attainment. These findings add to the evidence from Vietnam and indicate that wars inflict costs on the health of human populations that last longer than those relating to economic growth and welfare.

Keywords

disability health post-conflict Vietnam war

Introduction

Wars and armed conflicts are detrimental to the development of nations. While wars between states today are far less common than they were in the past, and civil wars are declining in number, the effects of past conflicts on exposed populations can last for many decades after the conflicts have ceased (Minoiu & Shemyakina, 2014; World Bank, 2011). This may be particularly the case for human capital accumulation, and health in particular. Attention is often focused on the death toll of war, especially in the short run, but war also has long-lasting impacts on the health of survivors and of people born after war ends. Wars inflict injury and trauma through direct exposure to violence and weapons. However, the indirect or ‘hidden’ effects of war on health may also be substantial and not surface until generations later (Ghobarah, Huth & Russett, 2003; Ugalde, Richards & Zwi, 1999). This may occur, for example, through the destruction or depletion of health care infrastructure and services; the reduction of arable land and nutrition; disruption to education and livelihoods; and contamination of the environment. Empirical evidence on the extent to which conflict impacts the long-term health of exposed populations, and in particular the channels through which this occurs, is limited, however, and has important implications for the development of policies to stem the consequences of conflict.

Economic growth theory predicts that war should have only temporary effects on the economy (Miguel & Roland, 2011). A transitory increase in capital accumulation and growth in war-affected areas is predicted until the steady state is reached. Macro-level studies in Japan, Germany, and Vietnam show that countries return to their steady state after 20–25 years (Brakman, Garretsen & Schramm, 2004; Davis & Weinstein, 2002; Miguel & Roland, 2011).¹ Similar postwar recovery patterns are predicted for human capital, including health. The effects of warfare on health may be overcome in the long term by postwar investments in public health care and other infrastructure, as well as via accumulation of human capital through investment in health inputs and behaviors (Strauss & Thomas, 2008). However, there are multiple hypotheses on why the health effects of war may be pervasive. For instance, wars place pressures on public resources towards economic and military reconstruction rather than the rebuilding of health systems; exposure to biological and chemical agents of warfare can cause delayed and intergenerational health effects; poor conditions experienced during pregnancy and early childhood may have irreversible consequences for health (Barker, 1992; Ghobarah, Huth & Russett, 2003; Ugalde, Richards & Zwi, 1999).

It is difficult to characterize the war in Vietnam, as it was both a civil and interstate conflict involving a variety of actors over time. The most intense period was marked by the military intervention of the United States and allied forces over the period 1964–75. The cornerstone of the US military strategy was an extensive bombing campaign, described as the most intense aerial bombing episode in history (Dell & Querubin, 2018; Kocher, Pepinsky & Kalyvas, 2011). Over six million tons of bombs and other ordnance were dropped in the former Indochina region at a weight four times greater than in Germany during World War II (Clodfelter, 1995). The loss of Vietnamese lives during the ‘American war’ period was immense, with estimates ranging from 1 million to 2 million (Hirschman, Preston & Vu, 1995; Obermeyer, Murray & Gakidou, 2008). Millions more were thought to be wounded (Kalyvas & Kocher, 2009). However, the long-term impacts of the war on the health of the Vietnamese population remain only partially documented. This question is particularly salient for Vietnam as Miguel & Roland (2011) found no significant long-term impact of US bombing on local poverty rates and other measures of economic development in Vietnam. Has Vietnam also overcome non-economic effects of the conflict with respect to health?

This article answers this question by estimating the impact of US bombing over the period 1965–75 on disability prevalence in Vietnam four decades later. Drawing upon a combination of national census and US air force data and an estimation strategy which instruments geographical bombing exposure with distance from the 17th parallel demilitarized zone, this article shows a significant link between wartime bombing and district-level disability prevalence approximately 40 years after the war. Highest effects are observed for severe disability and for persons aged around 40 years who were born in the years experiencing the highest density of bombing. Taken together with the work of Miguel & Roland (2011), our findings suggest that wars inflict external costs on the health of human populations that last longer than those relating to economic growth and welfare.

This article makes several noteworthy contributions. First, we contribute to the determinants of disability literature. Approximately 15% of the world’s population, more than a billion people, are estimated to be living with a disability (World Health Organization & World Bank, 2011). According to contemporary international classification, disability is understood as a dynamic interaction between health conditions and contextual factors, both personal and environmental (World Health Organization, 2001). For some environmental factors, such as those associated with malnutrition and poverty, the impact on health conditions associated with disability is well established (World Health Organization & World Bank, 2011). Major environmental changes, such as those caused by natural disasters or armed conflict, and their links to disability are less well understood. Ghobarah, Huth & Russett (2003) estimate the long-term effects of civil war on disability. However, the measure of disability – disability-adjusted life years (DALYs) – is controversial and the methods of analysis do not address the non-random nature of warfare exposure.² In this article we adopt a measure of disability consistent with the classification of the World Health Organization (2001), and a methodology which enables us to address unobserved heterogeneity.

Second, we contribute to the international evidence on the long-lasting effects of war on health. In line with Barker’s (1992) hypothesis linking early childhood environment and health outcomes later in life, most of this literature focuses on effects among affected children. Exposure to armed conflict in utero or in early childhood or adolescent years is associated with substantially lower height or height-for-age scores (a proxy for long-run nutritional status and health) in later life in multiple contexts (Akresh, Lucchetti & Thirumurthy, 2012; Bundervoet, Verwimp & Akresh, 2009; Grimard & Laszlo, 2014; Minoiu & Shemyakina, 2014). Exposure to armed conflict in childhood has also been linked to long-lasting detrimental effects on mental health, self-rated health satisfaction, BMI, and chronic health conditions such as stroke, hypertension, diabetes, and cardiovascular disease (Akbulut-Yuksel, 2014, 2017; Singhal, 2018).

This article adds to this literature by quantifying the long-term effects of conflict on disability. Specifically, we use a measure of disability developed by the United Nations’ Washington Group on Disability Statistics (WG hereafter), known as the WG Short-Set Questionnaire.³ The measure does not identify particular health conditions or impairments but instead captures the impact of a combination of factors (health included) on basic functional abilities (Madans, Loeb & Altman, 2010; Madans, Loeb & Eide, 2017). As such, the measure may be less susceptible to measurement error associated with other chronic health and disability measures. We estimate effects across the all-age population as well as heterogenous effects by birth cohort to provide a more complete picture of the population health effects of bombing.

Third, we contribute to the literature on the health effects of the 1965–75 conflict on the Vietnamese population. Most studies find that the war in Vietnam had a negative effect on health using a wide range of indicators for adults: functional and mental functioning (Singhal, 2018; Teerawichitchainan & Korinek, 2012); and for children: congenital impairments (Le et al., 1990), mortality (Hirschman, Preston & Vu, 1995; Savitz et al., 1993), infant neurodevelopment (Pham et al., 2013).⁴ Closest to our article is a study that estimates the long-term effects of bombing on mental health outcomes (Singhal, 2018). Using the 2016 wave of the Vietnam Access to Resources Survey and the same IV identification strategy, Singhal (2018) finds a substantial effect of early-life exposure to bombing on the likelihood of severe mental stress in adulthood. The survey, and representativeness of results, is restricted to rural households in 12 provinces.

Our article makes several contributions to the Vietnam literature. First, existing articles focus on subsamples of the Vietnam population that are not nationally representative. Second, this article is the first to use an internationally tested measure as an indicator for disability. Third, the vast majority of articles do not adopt an identification strategy to address unobserved heterogeneity. Drawing upon national census data, a standardized international health measure, and an instrumental variable approach, we isolate the impact of the war on disability prevalence at the district level 40 years after the war. In addition, we add to understanding on the mechanisms through which the war impacted on the long-term health of the Vietnamese people through the compilation of public healthcare data and improved or updated household survey data.

Conceptual framework

To more formally build our hypothesis about the long-term impact of bombing on disability in Vietnam, we draw upon general conceptual frameworks linking exposure to conflict and long-term health (Ghobarah, Huth & Russett, 2003; Ugalde, Richards & Zwi, 1999). Broadly, conflict gives rise to a range of conditions that increase the risk of disease and injury whose longer term consequences are exacerbated by other macro-level influences of conflict on public health (Ghobarah, Huth & Russett, 2003).⁵ Following Ugalde, Richards & Zwi (1999), we focus on several common conditions linking exposure to conflict and adverse consequences for health as relates to the war and disability in Vietnam.

The destruction and deterioration of healthcare infrastructure and services due to bombing could hasten the sequela of injuries and health conditions into long-term impairments and disability. While no link could be found between bombing and depleted public healthcare provision during the active war period in Vietnam (Dell & Querubin, 2018), economic difficulties and limited investments in public health infrastructure in the postwar period are well documented (Segall et al., 2002). While the network of community health stations was extensive, resources were not made available in a timely fashion to meet health worker salaries and the provision of drugs (Ensor & San, 1996). The resulting impact on the quality of services led to a series of health care reforms in the late 1980s. The introduction of user fees at public health facilities led to a decline in the use of formal health care and delays in treatment among the poor and rural population (Ensor & San, 1996). Given that the majority of bombing was in rural areas, it is possible that health conditions acquired as a consequence of the war deteriorated into functional difficulties or their difficulties increased in degree. Alternatively, it is possible that more heavily bombed areas experienced disproportionate levels of public infrastructure and health service investments in the postwar period such that the average long-term impact of bombing on disability has been reduced. Though not disaggregated by health, postwar state investments are positively correlated with bombing intensity (Miguel & Roland, 2011).⁶ Through the compilation of detailed provincial-level healthcare information on the number of hospitals, hospital beds, and doctors, we assess the extent to which bombing impacted on the long-run provision of healthcare services.

Bombing may induce negative impacts on nutritional status which, in turn, leads to long-term health conditions and disability. The reduction in food availability in conflict situations is often accompanied by deficiencies in protein, calories, and micronutrients which can cause severe malnutrition (Ugalde, Richards & Zwi, 1999). Malnutrition and poor living conditions in the fetal stage and early childhood is associated with a range of physical and mental health conditions in later life (Almond & Currie, 2011).⁷ It is estimated that 10% of sorties supported US allied troop ground operations and the remainder targeted enemy supply lines and insurgents (Dell & Querubin, 2018; Thayer, 1975). Targeting of physical and transportation infrastructure reduces food distribution whereas extensive aerial bombing and unexploded ordnance reduces the availability of arable land for food production. Loss of harvest and business activities reduces household economic resources for health production. Indeed, Singhal (2018) finds that exposure to bombing in early childhood in Vietnam significantly reduced adult height-for-age (a strong proxy for socio-economic conditions and nutrition in childhood) by 0.026 standard deviations or approximately 0.7 cm. However, the sample upon which the Singhal (2018) study was carried out is limited in size and geographical representativeness. We examine long-run effects of bombing on adult height using the Vietnam National Health Survey, a large nationally representative sample with extensive sampling in the central region where the heaviest bombing took place.

Conditions of conflict pose a substantial threat to economic resources which may impact on disability through its impact on nutritional status or through any number of other mediums, such as healthcare utilization or educational attainment (World Health Organization & World Bank, 2011). Miguel & Roland (2011) find no significant long-run effect of bombing on poverty rates in Vietnam, consistent with the finding of Singhal (2018) on weight-for-age (a proxy for current socio-economic conditions). We examine the impact of bombing on per capita consumption and poverty using updated household living standards and census data approximately 40 years after the war compared to Miguel & Roland’s (2011) estimates based on data reflecting impact 20–25 years after the war.

Access to education is an important and well-established input to health production over the life course which is commonly disrupted during violent conflict (Strauss & Thomas, 2008). There exists a relatively large literature establishing a causal link between conflict and reduced human capital among exposed populations (e.g. Akbulut-Yuksel, 2014; Inchino & Winter-Ebmer, 2004). In the context of Vietnam, Dell & Querubin (2018) find that bombing reduced access to primary school education by 16 percentage points in hamlets during the war time. Miguel & Roland (2011), however, find no long-term impact of bombing on literacy using 1999 census data. We add to this literature by estimating longer-run impacts on an alternate human capital measure (years of schooling completed) in 2009.

Exposure to biological and chemical agents used in warfare is associated with delayed and intergenerational health effects (Ugalde, Richards & Zwi, 1999). The conflict in Vietnam witnessed the first full-scale use of herbicides in warfare.⁸ Approximately two-thirds of the herbicides sprayed contained a highly toxic irremovable chemical, commonly known as dioxin, which is capable of concentrating in the body of animal and humans (Stellman et al., 2003). The US Institute of Medicine has concluded that a sufficient or suggestive link exists between exposure to dioxin contained in US military herbicides and a range of cancers and chronic health conditions, including spina bifida in the offspring of exposed people (Institute of Medicine, 2014). Multiple studies have attempted to estimate the health effects associated with military herbicide exposure on the Vietnamese population (Do, 2009; Le et al., 1990; Le & Johansson, 2001; Pham et al., 2013). For various reasons, some of which are mentioned above, the full extent of the population health effects are not known (Palmer, 2007). The primary purpose of the use of herbicides was as a defoliant to remove forest cover to the enemy, but it also aimed to destroy ‘unfriendly’ food crops and supplies (Stellman et al., 2003). Therefore, it is possible that the health effects of the herbicides occur through direct exposure to the dioxin or through indirect effects associated with nutrition. Unfortunately, our dataset does not contain information on the level of herbicide spraying so we are unable to formally test this mechanism or control for it in regressions. However, to the extent that bombing is correlated with herbicide spraying, our bombing measure will capture the effects of herbicides. One of the primary purposes of the spraying operation was to make bombing targets more visible, which suggests a degree of correlation (Stellman et al., 2003). Spraying maps resemble those of bombing maps with the highest levels of spraying taking place in the central province region surrounding the 17th parallel demilitarized zone and in regions where the ‘Ho Chi Minh Trail’ entered South Vietnam (Stellman et al., 2003).

Our bombing measure will also pick up delayed health effects associated with unexploded ordnance (UXO) and landmines. It is estimated that 10% of ordnance did not detonate on impact (Tran, Le & Husum, 2012). Due to the scale of the problem, the clearing of ordnance has been slow going with an estimated 19% of land still polluted by UXO and landmines (Tran, Le & Husum, 2012; Tuoitre, 2018).⁹ There exists an obvious correlation between bombing and UXO, and US and allied troop movements were concentrated in the regions with the highest level of bombing, which suggests a degree of correlation between bombing and the planting of landmines. To the extent that bombing is correlated with other weapons of warfare, our bombing measure is a proxy measure for overall war intensity and will pick up intergenerational effects on human health associated with these weapons. Our analysis extends to persons born before and after the cessation of bombing to capture the indirect effects associated with the above mechanisms as well as ongoing exposure to weapons.

Data and descriptive statistics

Data and variable definitions

This study draws upon two unique datasets for the main analysis of war intensity on disability as well as several other data sources for the analysis of mechanisms. War intensity is proxied by bombing data obtained from a database assembled by the US Defense Security Cooperation Agency.¹⁰ Specifically, our measure of bombing intensity is defined as the total number of bombs, missiles, and rockets per km² dropped by US and allied forces over the ten-year period 1965–75 (hereafter bombs, missiles, and rockets per km²). Bombs were the most common ordnance category and are highly correlated with other ordnance categories, for example cannon artillery and ammunition (Miguel & Roland, 2011). As previously discussed, bombing was central to the US military strategy and is correlated with other weapons, and thus represents a suitable proxy for war intensity.

To measure disability prevalence at the district level, we use the 15% sample of the 2009 Vietnam Population & Housing Census (VPHC). The sample is representative at the district level and contains approximately 3.7 million households and 14.2 million individuals. The survey contains a measure of disability recommended for use in national censuses and household surveys known as the WG Short-Set Questionnaire.¹¹ The 2009 VPHC included four of the six functional difficulty questions of the recommended short-set questionnaire: seeing, hearing, walking, and remembering. The four domains are considered the most essential in determining disability status from census data (Washington Group on Disability Statistics, 2017). This study defines persons with disabilities according to thresholds of functional difficulty as follows: (1) Persons with disability: persons experiencing difficulty of any degree (some difficulty, a lot of difficulty or unable to do) in at least one functional domain; (2) Persons with severe disability: persons experiencing a lot of difficulty or unable to do in at least one functional domain.

To analyze potential pathways from bombing to long-term impact on disability we draw upon several other data sources. We draw upon anthropometric height data contained in the 2001–02 Vietnam National Health Survey as a latent stock measure of health and marker for nutritional environment in childhood (Case & Paxson, 2010; Strauss & Thomas, 2008). We draw upon household consumption data contained in the 2010 Vietnam Household Living Standards Survey to construct measures of district-level per capita consumption and poverty rates. Measures were computed using the small area estimation method with the 2009 VPHS, as described in Lanjouw, Marra & Nguyen (2017). We draw upon data on number of years of schooling completed (for persons aged 22 years and above) contained in the 2009 VPHS to compute a measure of district-level human capital attainment. Finally, we use public health information contained in the 2009 Statistical Yearbook of Vietnam to construct

Table I.

District-level disability prevalence

	Districts by quintiles of bombs, missiles, rockets per km²
	Lowest	Near lowest	Middle	Near highest	Highest	Total
Disability prevalence by functioning domain
Seeing	5.16	5.21	4.88	4.76	5.79	5.15
Hearing	3.43	3.11	3.13	2.83	3.58	3.20
Walking	3.89	3.69	3.72	3.45	4.25	3.79
Remembering	3.68	3.52	3.58	3.27	4.06	3.61
Any domain	8.12	7.90	7.65	7.23	8.75	7.91
Severe disability prevalence by functioning domain
Seeing	0.63	0.61	0.62	0.61	0.77	0.65
Hearing	0.65	0.61	0.63	0.58	0.74	0.64
Walking	0.90	0.88	0.91	0.89	1.09	0.93
Remembering	0.80	0.80	0.84	0.78	0.97	0.84
Any domain	1.70	1.63	1.71	1.62	1.99	1.73

measures on the number of hospitals, patient beds, and doctors at the provincial level.

Descriptive statistics

Table I presents the average percentage of people aged above five with disabilities in each of the four functional domains and overall at district level. The percentage of people with disability is around 8%. The average rates of persons experiencing difficulty in seeing, hearing, walking, and remembering at the district level are 5.2%, 3.2%, 3.8%, and 3.6%, respectively. The prevalence of severe disability is 1.7%. For both the general and severe disability measures, prevalence rates are higher in districts with higher bombing density (Table I) and are significantly higher among persons born before versus after the war (Table II); for example, on average 15% of the district population born before 1976 have a disability compared to 1.8% born since 1976. Figure 1 illustrates significant heterogeneity in the density of bombing and disability prevalence across districts in Vietnam.

Empirical strategy

To estimate the long-run impact of bombing on disability in Vietnam we start with simple ordinary least squares (OLS) regression where we assume that disability prevalence at the district level is a reduced-function of the density of bombs and other control variables as follows:

Table II.

District-level disability prevalence for people born before and since 1976

	Districts by quintiles of bombs, missiles, rockets per km²
	Lowest	Near lowest	Middle	Near highest	Highest	Total
People born before 1976
Disability	15.57	14.80	14.44	14.12	16.24	15.00
Severe disability	3.02	2.85	2.97	2.91	3.43	3.03
People born since 1976
Disability	1.84	1.85	1.81	1.58	1.92	1.79
Severe disability	0.59	0.57	0.62	0.56	0.69	0.61

L o g (d i s a b i l i t y_{i}) = α + L o g (b o m b_{i}) β + X^{'} θ + u_{i}

where $L o g (d i s a b i l i t y_{i})$ is the log of the rate of disability prevalence in district i; $L o g (b o m b_{i})$ is the log of bomb density in district i, which is measured by the total number of bombs, missiles, and rockets per km² dropped by US and allied forces in the district over the period 1965–75; X′ is a vector of exogenous control variables; and u_i denotes unobserved variables.¹² Control variables include those that are not affected by the treatment variable of bomb density: district area and elevation, the share of urban population, district capital and Northern region dummy, and distance to the major cities Da Nang, Hanoi, and Ho Chi Minh city (Angrist & Pischke, 2008; Heckman, Lalonde & Smith, 1999). Standard errors are clustered at both district and province levels using a multilevel clustering technique to address within- and between-correlation of errors at both the district and provincial levels simultaneously (Cameron, Gelbach & Miller, 2011).

Since the US bombing was not random, it can be correlated with omitted variables and as a result OLS estimators can be biased. A standard econometric method to correct this bias is through an instrumental variable approach (Angrist, Imbens & Rubin, 1996). An instrument must be highly correlated with the bombing variable but not the error term u_i. We follow the approach of Miguel & Roland (2011) and use the distance from the center of each district to the 17th parallel north latitude as an instrument of bombing density. The 17th parallel north latitude was set in 1954 as the border between the former northern and southern Vietnamese territories. Since this was a border, it was heavily bombed and targeted by different types of weapons. The first stage is expressed as follows:

L o g (b o m b_{i}) = α + L o g (d i s t a n c e_17 l a t_{i}) β + X^{'} θ + ε_{i}

where $L o g (d i s t a n c e_17 l a t_{i})$ is the log of the distance from the center of district i to the 17th parallel north latitude. An important point in the identification strategy is that the determination of the 17th parallel was the product of negotiations between the United States and the Soviet Union in the context of the Cold War (Miguel & Roland, 2011). To the extent that the border was determined by factors external to Vietnam (rather than local geographical or socio-economic factors which may be jointly correlated with disability status) and is correlated with the intensity of bombing, proximity to the former north–south border can be viewed as a natural experiment with which to estimate the impact of bombing on disability prevalence.

Our identification strategy nonetheless remains open to potential bias. One particular concern relates to migration. If migration patterns during or after the war are jointly correlated with distance to the former border and the prevalence of disability then our results are biased. Unfortunately, no reliable record of migration during or in the immediate postwar period exists. However, no statistical relationship was found between bombing and population growth over the active (1969–73) and postwar (1985–2000) periods (Dell & Querubin, 2018; Miguel & Roland, 2011). Furthermore, bombing did not have a consistent effect on the proportion of people not born in their current village of residence in 1997–98 (Miguel & Roland, 2011). The qualitative literature suggests that forced urbanization happened to a lesser extent than expected and that the vast majority of households remained or returned shortly after bombing (Appy, 2015; Tirman, 2011). While we cannot rule out selective migration on the basis of disability status in heavily bombed regions, these findings suggest that any potential bias is small.

Results

We start with double-log OLS regressions of disability on bombing intensity. In all models, the coefficient of the bomb variable is positive and statistically significant at standard levels of significance (Tables III and IV).

Figure 1.

Bomb density and disability rate at the district level

Table III.

OLS regression of disability prevalence rates

Explanatory variables	Seeing	Hearing	Walking	Remembering	Any domain
Log of bombs, missiles, rockets per km²	0.0170^†	0.0124^†	0.0129^†	0.0223**	0.0149*
	(0.0088)	(0.0064)	(0.0068)	(0.0071)	(0.0060)
Log of area of district	–0.0297	–0.0700**	–0.0372	–0.0476^†	–0.0344
	(0.0323)	(0.0261)	(0.0276)	(0.0271)	(0.0239)
Log of mean elevation	–0.0237*	–0.0017	–0.0326**	–0.0159	–0.0141
	(0.0116)	(0.0098)	(0.0097)	(0.0103)	(0.0089)
Northern (yes=1, no=0)	0.2355**	0.4126**	0.3605**	0.3548**	0.3333**
	(0.0393)	(0.0335)	(0.0331)	(0.0357)	(0.0302)
Capital district of province (yes=1, no=0)	–0.0609	0.0123	0.0241	0.0446	0.0347
	(0.0780)	(0.0536)	(0.0544)	(0.0559)	(0.0486)
Log of distance to closest cities: Hanoi, Da Nang, HCMC	0.0597*	0.0626**	0.0599**	0.0494*	0.0482**
	(0.0262)	(0.0202)	(0.0219)	(0.0206)	(0.0184)
Share of urban population	–0.2972*	–0.5798**	–0.4377**	–0.5326**	–0.4581**
	(0.1195)	(0.0893)	(0.0831)	(0.0882)	(0.0777)
Constant	–3.4176**	–3.6198**	–3.7433**	–3.6151**	1.4648**
	(0.2173)	(0.1943)	(0.1932)	(0.1932)	(0.1726)
Observations	612	612	612	612	612
R-squared	0.124	0.399	0.298	0.303	0.333

**p < 0.01, *p < 0.05, ^†p < 0.1. Robust clustered standard errors in parentheses.

The coefficients are higher in magnitude and level of significance for severe disability compared to overall disability. A 10% increase in bombing density leads to relative increases of 0.30% in severe disability prevalence and 0.15% in overall disability prevalence.

Since the OLS estimator could be biased, we primarily rely on the IV regression. The first stage regression shows a strong correlation between the instrument and bomb density (Online appendix Table A.II). Districts further from the 17th latitude are less likely to receive

Table IV.

OLS regression of severe disability prevalence rates

	Severe	Severe	Severe	Severe	Severe
Explanatory variables	Seeing	Hearing	Walking	Remembering	Any domain
Log of bombs, missiles, rockets per km²	0.0319**	0.0261**	0.0293**	0.0390**	0.0301**
	(0.0093)	(0.0077)	(0.0079)	(0.0076)	–0.0069
Log of area of district	–0.0375	–0.0567^†	–0.0417	–0.0635*	–0.0471^†
	(0.0356)	(0.0306)	(0.0313)	(0.0303)	(0.0275)
Log of mean elevation	–0.0195	–0.0041	–0.0358**	–0.018	–0.0181^†
	(0.0127)	(0.0114)	(0.0110)	(0.0112)	(0.0101)
Northern (yes=1, no=0)	0.3847**	0.4599**	0.3854**	0.4256**	0.3849**
	(0.0425)	(0.0384)	(0.0365)	(0.0371)	(0.0341)
Capital district of province (yes=1, no=0)	–0.1035	–0.0164	0.0206	0.0266	0.0368
	(0.0806)	(0.0642)	(0.0611)	(0.0607)	(0.0556)
Log of distance to closest cities: Hanoi, Da Nang, HCMC	0.0553*	0.0455^†	0.0451^†	0.0299	0.0378^†
	(0.0281)	(0.0242)	(0.0245)	(0.0220)	(0.0209)
Share of urban population	–0.4266**	–0.5542**	–0.3893**	–0.4990**	–0.4488**
	(0.1213)	(0.1014)	(0.0929)	(0.0906)	(0.0840)
Constant	–0.9243**	–0.7264**	–0.5511*	–0.4153*	0.0918
	(0.2459)	(0.2184)	(0.2139)	(0.2112)	(0.1962)
Observations	612	612	612	612	612
R-squared	0.220	0.344	0.256	0.318	0.314

**p < 0.01, *p < 0.05, ^†p < 0.1. Robust clustered standard errors in parentheses.

bombs, missiles, and rockets with a t-statistic equal to 10.¹³ Districts further away from major cities were more likely to be bombed, though the effect is only mildly significant, whereas those that are located in northern regions or are higher in altitude experienced less bombing. There is no robust correlation between the level of bombing and district size, district capital, and share of the urban population.

The IV estimates for the general disability measure show a highly significant effect of natural log of bomb density on the natural log of disability prevalence (Table V). A 10% increase in ordnance leads on average to a 0.61% relative increase in the overall proportion of people with disability. Similarly, if the ordnance dropped increases by 10% the proportions of people experiencing difficulty in seeing, in hearing, in walking, and in remembering increase in relative terms by 0.50%, 0.52%, 0.52%, and 0.80%, respectively.

As illustrated in Figure 1, there was considerable variation in bombing intensity across districts. To better gauge the magnitude of the effect across districts, we disaggregate districts in quintiles by means of logged bombing intensity. Districts in the top quintile have a bombing intensity that is 406% that of districts in the bottom quintile. This translates into a 25% (406 x 0.061) relative increase in the rate of disability attributable to bombing in a high- versus low-bombed district. Given a mean disability prevalence of 8.75% in top quintile districts (Table I), this infers a 1.75 absolute percentage point decrease in disability prevalence if heavily bombed districts experienced a bombing intensity similar to the least bombed districts.¹⁴

A 10% increase in ordnance leads to a 0.88% relative increase in the proportion of people with severe disability (Table VI). Using a similar calculation as above, bombing accounts for a 36% relative increase in the prevalence of severe disability in a high- versus low-bombed district, equating to a 0.53 percentage point decrease in the rate of severe disability had heavily bombed districts experienced a level of bombing equivalent to low bombed district.¹⁵ Results illustrate the higher relative long-term impact of bombing on levels of severe disability compared with overall disability.

To examine whether the impact of bombing varies across age cohorts, we ran regressions of the log of the district disability rate computed for different age cohorts on the log of bomb density using the preferred IV

Table V.

IV regression of disability prevalence rates

Explanatory variables	Seeing	Hearing	Walking	Remembering	Any domain
Log of bombs, missiles, rockets per km²	0.0500**	0.0523**	0.0516**	0.0800**	0.0606**
	(0.0131)	(0.0103)	(0.0109)	(0.0119)	(0.0095)
Log of area of district	–0.0307	–0.0711**	–0.0383	–0.0493^†	–0.0357
	(0.0322)	(0.0266)	(0.0280)	(0.0286)	(0.0248)
Log of mean elevation	–0.0161	0.0075	–0.0236*	–0.0026	–0.0035
	(0.0117)	(0.0100)	(0.0099)	(0.0108)	(0.0091)
Northern (yes=1, no=0)	0.2522**	0.4328**	0.3801**	0.3841**	0.3565**
	(0.0394)	(0.0340)	(0.0340)	(0.0371)	(0.0312)
Capital district of province (yes=1, no=0)	–0.066	0.0062	0.0181	0.0357	0.0276
	(0.0796)	(0.0576)	(0.0586)	(0.0630)	(0.0543)
Log of distance to closest cities: Hanoi, Da Nang, HCMC	0.0555*	0.0575**	0.0550*	0.0421^†	0.0424*
	(0.0256)	(0.0204)	(0.0221)	(0.0216)	(0.0191)
Share of urban population	–0.3170**	–0.6038**	–0.4610**	–0.5674**	–0.4856**
	(0.1185)	(0.0904)	(0.0845)	(0.0920)	(0.0805)
Constant	–3.4351**	–3.6409**	–3.7637**	–3.6456**	1.4407**
	(0.2202)	(0.2009)	(0.2000)	(0.2060)	(0.1818)
Observations	612	612	612	612	612
R-squared	0.099	0.357	0.254	0.214	0.260

**p < 0.01, *p < 0.05, ^†p < 0.1. Robust clustered standard errors in parentheses.

Table VI.

IV regression of severe disability prevalence rates

	Severe	Severe	Severe	Severe	Severe
Explanatory variables	Seeing	Hearing	Walking	Remembering	Any domain
Log of bombs, missiles, rockets per km²	0.0904**	0.0791**	0.0784**	0.1016**	0.0881**
	(0.0149)	(0.0126)	(0.0139)	(0.0125)	(0.0117)
Log of area of district	–0.0392	–0.0582^†	–0.0431	–0.0653*	–0.0488^†
	(0.0361)	(0.0315)	(0.0322)	(0.0317)	(0.0290)
Log of mean elevation	–0.006	0.0081	–0.0245*	–0.0035	–0.0047
	(0.0130)	(0.0116)	(0.0116)	(0.0116)	(0.0105)
Northern (yes=1, no=0)	0.4143**	0.4868**	0.4103**	0.4574**	0.4143**
	(0.0439)	(0.0400)	(0.0382)	(0.0393)	(0.0363)
Capital district of province (yes=1, no=0)	–0.1125	–0.0245	0.013	0.0169	0.0279
	(0.0866)	(0.0701)	(0.0672)	(0.0690)	(0.0642)
Log of distance to closest cities: Hanoi, Da Nang, HCMC	0.0478^†	0.0388	0.0388	0.0219	0.0304
	(0.0284)	(0.0246)	(0.0249)	(0.0231)	(0.0218)
Share of urban population	–0.4618**	–0.5861**	–0.4189**	–0.5367**	–0.4838**
	(0.1242)	(0.1046)	(0.0966)	(0.0955)	(0.0890)
Constant	–0.9553**	–0.7544**	–0.5770*	–0.4484*	0.0611
	(0.2548)	(0.2289)	(0.2241)	(0.2247)	(0.2097)
Observations	612	612	612	612	612
R-squared	0.159	0.287	0.197	0.226	0.219

**p < 0.01, *p < 0.05, ^†p < 0.1. Robust clustered standard errors in parentheses.

specification.¹⁶ Figure 2 presents the estimated effect of log of bomb density on log of the proportion of people with disability at different ages. There is a clear inverted-U shape relationship between the impact of bombing on disability and age. Highest effects occur for people aged around 40 years, the very group born in the years experiencing the highest density of bombing during the war. However, the impact of bombing on disability

Figure 2.

The estimated effect of log of bomb density on log of the proportion of people with disability by age cohort

continues for a long period of time. People born after the war are still impacted by the bombing, although the effect tends to be smaller than for people born before or during the war.

Mechanisms

As outlined in the conceptual framework, there are multiple pathways through which bombing can impact on disability in the long run. In Table VII, we present results on the long-term impact of bombing on several channels as relating to nutritional status (proxied by height), living standards, human capital, and public healthcare infrastructure and services using our preferred IV specification.¹⁷ As shown in the first column, bombing had a negative effect on recorded height in 2001–02. A 10% increase in bombing leads on average to a 0.035 cm reduction in height. The average effect is small but significant at the 1% level. For heavily bombed districts the effect is much larger. Districts in the top quintile experienced a bombing intensity 406% greater than districts in the bottom quintile. This translates into an average 1.42 cm (406 x 0.0035) reduction in height for people living in a high- versus low-bombed district. Similarly, we find a significant negative effect of bombing on average years of schooling completed for persons 22 years of age and above at district level in the year 2009 (column 4). A 10% increase in bombing leads on average to a 0.01 year reduction in schooling. This equates to an average 0.41 year (406 x 0.001) reduction in schooling for people living in a high- versus low-bombed district and the rate of poverty at district level, and the number of hospitals, patient beds, and doctors at provincial level in 2009.

Discussion

Overall, our results contribute to the emerging literature on the long-lasting impacts of war on health, measured for the first time through questions from an internationally tested and comparable disability questionnaire. Contrary to an earlier study on the long-run economic impact of the bombing in Vietnam (Miguel & Roland, 2011), we find a positive effect on non-economic outcomes as measured by disability. As expected, we find a relatively larger effect on severe disability prevalence and among persons born before or during the conflict. Effects decline with age for cohorts born before the war, which may reflect excess mortality in these age groups compared to younger ones. We also find significant, albeit smaller, disability effects among persons born after the cessation of bombing, which may be explained by any number of indirect mechanisms and continued exposure to UXO, landmines, and sprayed chemicals that were correlated with the level of bombing. Our results are consistent with findings relating to long-term effects of bombing on mental health outcomes in Vietnam (Singhal, 2018), as well as studies which find an association between war exposure and a range of health outcomes among Vietnamese people born before and after the end of the war (Le et al., 1990; Le & Johansson, 2001; Ngo et al., 2006; Pham et al., 2013).

It is difficult to disentangle the pathways from bombing to long-run disability in the absence of detailed data in the post-conflict period. We examine the long-term impact of bombing on several common channels as identified in our conceptual framework. Consistent with Miguel & Roland (2011), we find no significant effects of bombing on long-term poverty rates and consumption per capita. However, consistent with Singhal (2018), we do find a negative long-term effect of bombing on height which is a strong proxy for socio-economic conditions and nutrition in utero and early childhood (Case & Paxson, 2010; Strauss & Thomas, 2008). Poverty and a lack of essential dietary nutrients (such as iodine or folic acid) lead to the onset of a range of health conditions associated with disability (World Health Organization & World Bank, 2011). Consequently, adverse conditions created or compounded by

Table VII.

IV regression of other outcomes

Explanatory variables	Height	Log of per capita expenditure	Log of poverty rate	Years of schooling completed	Log of hospitals per one million people	Log of patient beds per one million people	Log of doctors per one million people
Log of bombs, missiles, rockets per km²	–0.3505**	–0.0077	–0.0051	–0.1011*	0.0123	0.0312	0.0324
Log of bombs, missiles, rockets per km²	(0.056)	(0.0098)	(0.0209)	(0.0485)	(0.0452)	(0.0441)	(0.0596)
Log of area of district	0.0380	–0.0771**	0.2090**	–0.5489**	0.0786	0.0745	–0.0403
	(0.086)	(0.0161)	(0.0519)	(0.0955)	(0.0955)	(0.0899)	(0.1026)
Log of mean elevation	–0.4562**	–0.0401**	0.0917**	–0.0337	–0.019	–0.0601^†	–0.0578
	(0.041)	(0.0068)	(0.0169)	(0.0356)	(0.0283)	(0.0330)	(0.0399)
Northern (yes=1, no=0)	–0.5453**	–0.0628**	0.0035	0.9509**	0.2498*	0.3287*	0.192
	(0.145)	(0.0208)	(0.0621)	(0.1127)	(0.1204)	(0.1496)	(0.1785)
Capital district of province (yes=1, no=0)	0.3985*	0.1527**	–0.1867	1.0599**
Capital district of province (yes=1, no=0)	(0.190)	(0.0350)	(0.1418)	(0.1782)
Log of distance to closest cities: Hanoi, Da Nang, or HCMC	–0.1699*	–0.1143**	0.4125**	–0.3929**
Log of distance to closest cities: Hanoi, Da Nang, or HCMC	(0.080)	(0.0135)	(0.0499)	(0.0665)
Share of urban population	0.6996**	0.3773**	–1.2241**	0.7245**	1.2495**	1.2455	1.4907^†
	(0.153)	(0.0538)	(0.1884)	(0.2687)	(0.4130)	(0.9808)	(0.8230)
Age	3.1049**
	(0.014)
Age²	–0.0344**
	(0.000)
Male (male=1, female=0)	6.6990**
	(0.083)
Constant	97.1858**	10.2646**	0.3979	11.1557**	1.3678^†	6.7613**	6.2281**
	(0.442)	(0.1197)	(0.4126)	(0.6336)	(0.7250)	(0.7212)	(0.7761)
Observations	137,543	612	612	612	63	63	63
R-squared	0.726	0.734	0.731	0.543	0.144	0.242	0.274

**p < 0.01, *p < 0.05, ^†p < 0.1. Robust clustered standard errors in parentheses.

the war likely contributed to the future ill-health and disability of Vietnamese people. As shown in Online appendix Table A.III, we observe larger long-term effects of bombing on height among individuals born before the end of the war versus those born after, consistent with patterns on disability (Online appendix Table A.I). Effects among those alive during the war may be explained by the substantial threats to economic resources, food supply, and production. The early postwar reunification period was characterized by widespread food shortages and economic difficulty that may have impacted on cohorts born in areas affected by the war differentially, for example through the reduction of arable land due to UXO and landmine contamination (Tran, Le & Husum, 2012). Our results add to a growing empirical literature on the negative impact of conflict on early living conditions and nutrition as a pathway to poor long-term health status (Akresh, Lucchetti & Thirumurthy, 2012; Bundervoet, Verwimp & Akresh, 2009; Grimard & Laszlo, 2014; Minoiu & Shemyakina, 2014).

We find a negative effect of bombing density on long-run human capital attainment as measured by years of schooling. As shown in Online appendix Table A.III, the effect is contained to people born before the end of the war with no significant effect observed for people born after 1975. The result suggests that human capital attainment as an important input to health production is a possible median through which bombing influenced disability over the long term in Vietnam. Our result is consistent with Dell & Querubin (2018), who find that bombing reduced access to primary school education in hamlets during the wartime, and is inconsistent with Miguel & Roland (2011), who find no impact of bombing on long-term literacy rates. More broadly, the result is consistent with other contexts where exposure to conflict is associated with depletion in human capital (Akbulut-Yuksel, 2014; Inchino & Winter-Ebmer, 2004).

We do not find any discernible long-run impact of bombing on measures of public health infrastructure and services. Similar to the economic story, this suggests that public resources may have been equalized across provinces with varying exposure to the war such that the impacts of war on the public health system have disappeared over time. There exists no apparent pathway from healthcare destruction or deterioration to disability over the long term at least. However, as previously discussed, the public health care system deteriorated significantly in the post-reunification period, which could hasten the sequela of injuries and health conditions into long-term impairments and disability. Unfortunately, there exist no data on provincial-level rehabilitation services. However, disability-specific health care services, including rehabilitation, and assistive devices were likely to have been very limited in the postwar period and remain in short supply today (Palmer et al., 2015). This all points to an enduring effect of war-related injuries and impairments on disability.

We do not have access to complete data on UXO, landmines, and herbicides so cannot test these mechanisms. However, there exists an obvious relationship with our ordnance measure and UXO, and the planting of landmines and spraying of herbicides is likely to be correlated with the density of bombing in the central and the southern regions. To this extent, our bombing measure will pick up long-term effects on disability associated with exposure to these weapons. According to the Vietnam National Mine Action Centre,¹⁸ since 1975 some 40,000 people have lost their lives and 60,000 persons have been injured due to UXO and mines (VOV, 2016). One-fifth of the injuries derive from six central region provinces where bombing was most concentrated. In Quang Tri province, the heaviest bombed province, the vast majority of injuries stemmed from cluster bombs dropped from aircraft and other ordnance, with less than 5% from landmines (Tran, Le & Husum, 2012). Our bombing measure may therefore capture the majority of disabling effects associated with exposure to UXO and landmines. To the extent that bombing and spraying patterns differed, further research is required to better establish the long-term population health effects associated with military herbicide exposure.

Our results taken together with those of Miguel & Roland (2011) suggest that postwar economic recovery patterns may be different from human capital patterns, as measured through disability prevalence, which is consistent with findings from postwar Germany (Akbulut-Yuksel, 2014, 2017). The overall magnitude of the effect of bombing on the long-term disability rate among the Vietnamese population is small though statistically significant. However, there exists large variation across districts reflecting variation in bombing intensity; a heavily bombed district has 25% more people with disabilities and 36% more people with severe disabilities due to experiencing a high rather than low bombing intensity during the war. Given that millions of people died during the war and in the four decades since, our findings are surprising and appear to reflect what may have in earlier years been higher rates of disability in heavily bombed areas. In addition, our disability measure is unable to accurately capture a range of mental health conditions commonly associated with warfare exposure, such as depression, anxiety, and post-traumatic stress disorder (Do & Iyer, 2012). We offer a partial and long-term estimate of the effect of war on health for the entire population of Vietnam, which is new in the literature.

Conclusion

Vietnam bore the brunt of the most intense and sustained aerial bombing campaign in military history. In this article, we examine the long-term effects on health as measured by an internationally standardized measure of disability. This article shows a significant link between wartime bombing and district level disability prevalence approximately 40 years after the war. Highest effects are observed for severe disability and for persons aged around 40 years who were born in the years experiencing the highest density of bombing. Smaller yet significant effects are found for persons born after the war. Disentangling the pathways to long-term disability effects is complex. Our results combined with a review of the theory and historical literature suggest that the likely pathway is through direct exposure to the bombing and other correlated ordnance and weapons including UXO, landmines, and dioxin, as well as through indirect mechanisms relating to poor nutritional environment, human capital depletion, and low quality public healthcare. Taken together with the earlier work of Miguel & Roland (2011) and Singhal (2018), our findings suggest that wars inflict costs on the health of human populations that last longer than those relating to economic growth and welfare. These findings from Vietnam underpin the importance of expediting the process of cleaning up the consequences of war, and in ensuring food security and adequate health and rehabilitation services for people in conflict-affected zones. The toll of warfare is often assessed in terms of the number of people killed. However, the long-term consequence of warfare on disability is significant and, as this article suggests, deserves closer attention.

Supplemental Material

Supplemental Material, JPR846545_Appendix - Long-lasting consequences of war on disability

Supplemental Material, JPR846545_Appendix for Long-lasting consequences of war on disability by Michael Palmer, Cuong Viet Nguyen, Sophie Mitra, Daniel Mont and Nora Ellen Groce in Journal of Peace Research

Footnotes

Replication data

The Online appendices, dataset, codebook and do-files for the empirical analysis in this article can be found at .

Acknowledgements

We gratefully acknowledge comments from the editor Gudrun Østby and three anonymous reviewers, Dennis Petrie, participants of the IRDES-Dauphine Workshop on Applied Health Economics and Policy Evaluation (Paris, France), the GREThA International Conference on Economic Development (Bordeaux, France), and the Australian Health Economics Society Conference (Hobart, Australia).

ORCID iD

Michael Palmer

Notes

References

Akbulut-Yuksel

Mevlude

(2014) Children of war: The long-run effects of large-scale physical destruction and warfare on children. Journal of Human Resources 49(3): 634–662.

Akbulut-Yuksel

Melvude

(2017) War during childhood: The long run effects of warfare on health. Journal of Health Economics 53(May): 117–130.

Akresh

Richard

Lucchetti

Leonardo

Thirumurthy

Harsha

(2012) Wars and child health: Evidence from the Eritrean–Ethiopian conflict. Journal of Development Economics 99(2): 330–340.

Almond

Douglas

Currie

Janet

(2011) Human capital development before age five. In: Ashenfelter

Orley

Card

David

(eds) Handbook of Labour Economics, Vol. 4B. San Diego, CA: Elsevier, 1315–1486.

Angrist

Joshua D

Pischke

Jorn-Steffan

(2008) Mostly Harmless Econometrics: An Empiricist’s Companion. Princeton, NJ: Princeton University Press.

Angrist

Joshua D

Imbens

Guido W

Rubin

Donald B

(1996) Indentification of causal effects using instrumental variables. Journal of the American Statistical Association 91(434): 444–455.

Appy

Christian G

(2015) American Reckoning: The Vietnam War and Our National Identity. New York: Penguin.

Barker

David James Purslove

(1990) The fetal and infant origins of adult disease. British Medical Journal 301(6761): 1111.

Brakman

Steven

Garretsen

Harry

Schramm

Marc

(2004) The strategic bombing of German cities during World War II and its impact on city growth. Journal of Economic Geography 4(2): 201–218.

10.

Bundervoet

Tom

Verwimp

Philip

Akresh

Richard

(2009) Health and civil war in rural Burundi. Journal of Human Resources 44(2): 536–563.

11.

Cameron

Colin

Gelbach

Jonah B

Miller

Douglas L

(2011) Robust inference with multi-way clustering. Journal of Business and Economic Statistics 29(2): 238–249.

12.

Case

Anne

Paxson

Christina

(2010) Causes and consequences of early life health. Demography 47(Suppl. 1): S65–S85.

13.

Clodfelter

Michael

(1995) Vietnam in Military Statistics: A History of the Indochina Wars, 1772–1991. Jefferson, NC: McFarland.

14.

Cragg

John G

Donald

Stephen G

(1993) Testing identifiability and specification in instrumental variable models. Econometric Theory 9(2): 222–240.

15.

Davis

Donald R

Weinstein

David E

(2002) Bones, bombs and break points: The geography of economic activity. American Economic Review 92(5): 1269–1289.

16.

Dell

Melissa

Querubin

Pablo

(2018) Nation building through foreign intervention: Evidence from discontinuities in military strategy. Quarterly Journal of Economics 133(2): 701–764.

17.

Quy-Toan

(2009) Agent Orange and the Prevalence of Cancer among the Vietnamese Population 30 Years after the End of the Vietnam War. Washington DC: World Bank.

18.

Quy-Toan

Iyer

Lakshmi

(2012) Mental health in the aftermath of conflict. In: Garfinkel

Michelle R

Skaperdas

Stergios

(eds) The Oxford Handbook of the Economics of Peace and Conflict. Oxford: Oxford University Press, 341–361.

19.

Ensor

Tim

San

Pham Bich

(1996) Access and payment for health care: The poor of northern Vietnam. International Journal of Health Planning and Management 11(1): 69–83.

20.

Ghobarah

Hazem Adam

Huth

Paul

Russett

Bruce

(2003) Civil wars kill and maim people – long after the shooting stops. American Journal of Political Science 97(2): 189–202.

21.

Grimard

Laszlo

(2014) Long-term effects of civil conflict on women’s health outcomes in Peru. World Development 54(February): 139–155.

22.

Heckman

James J

Lalonde

Robert J

Smith

Jeffrey A

(1999) The economics and econometrics of active labor market programs. In: Ashenfelter

Orley

Card

David

(eds) Handbook of Labour Economics. Amsterdam: Elsevier Science, 1865–2097.

23.

Hirschman

Charles

Preston

Samuel

Loi Manh

(1995) Vietnamese casualities during the American War: A new estimate. Population and Development Review 21(4): 783–812.

24.

Hoang

(2018) Ký ức buồn của những người suýt bỏ mạng, vướng lao lý vì mưu sinh bằng ‘nghề’ đào bomb [The sad memories of those who nearly lost their lives or went to jail due to the bomb mining ‘industry’]. Doi Song Phat Luat 18 December, no. 4.

25.

Inchino

Andrea

Winter-Ebmer

Rudolf

(2004) The long-run education cost of World War II. Journal of Labour Economics 22(1): 57–86.

26.

Institute of Medicine (2014) Veterans and Agent Orange. Washington, DC: National Academies.

27.

Kalyvas

Stathis N

Kocher

Matthew Adam

(2009) The dynamics of violence in Vietnam: An analysis of the hamlet evaluation system. Journal of Peace Research 46(3): 335–355.

28.

Kleibergen

Frank

Paap

Richard

(2006) Generalized reduced rank tests using the singular value decomposition. Journal of Econometrics 133(1): 97–126.

29.

Kocher

Matthew Adam

Pepinsky

Thomas B

Kalyvas

Stathis N

(2011) Aerial bombing and insurgency in the Vietnam War. American Journal of Political Science 55(2): 201–218.

30.

Lanjouw

Peter

Marra

Marleen

Nguyen

Cuong

(2017) Vietnam’s evolving poverty index map: Patterns and implications for policy. Social Indicators Research 133(1): 93–118.

31.

Dai Cao

Nguyen

Phuong Ngoc Thi

Thom Hong

Tran

Thuy Thi

Nguyen

Van Tu Thi

Cam Hong

Huynh

Chi Kim Thi

Thuy Bich

(1990) A comparison of infant mortality rates between two Vietnamese villages sprayed by defoliants in wartime and one unsprayed village. Chemosphere 20(7–9): 1005–1012.

32.

Tuyet Nham Thi

Johansson

Annika

(2001) Impact of chemical warfare with Agent Orange on women’s reproductive lives in Vietnam: A pilot study. Reproductive Health Matters 9(18): 156–164.

33.

Madans

Jennifer H

Loeb

Mitchell E

Altman

Barbara M

(2010) Measuring disability and monitoring the UN Convention on the Rights of Persons with Disabilities: The work of the Washington Group on Disability Statistics. BMC Public Health 11(S4): 1–8.

34.

Madans

Jennifer

Loeb

Mitchell

Eide

Arne H

(2017) Measuring disability and inclusion in relation to the 2030 agenda on sustainable development. Disability and the Global South 4(1): 1164–1179.

35.

Miguel

Edward

Roland

Gerard

(2011) The long-run impact of bombing Vietnam. Journal of Development Economics 96(1): 1–15.

36.

Minoiu

Camelia

Shemyakina

Olga N

(2014) Armed conflict, household victimization, and child health in Côte d’Ivoire. Journal of Development Economics 108(May): 237–255.

37.

Mont

Daniel

(2007) Measuring health and disability. Lancet 369(9573): 1658–1663.

38.

Ngo

Anh D

Taylor

Richard

Roberts

Christine L

Nguyen

Tuan V

(2006) Association between Agent Orange and birth defects: Systematic review and meta-analysis. International Journal of Epidemiology 35(5): 1220–1230.

39.

Obermeyer

Ziad

Murray

Christopher JL

Gakidou

Emmanuela

(2008) Fifty years of violent war deaths from Vietnam to Bosnia: Analysis of data from the World Health Survey Programme. British Medical Journal 336(7): 1482–1486.

40.

Palmer

Michael G

(2007) The case of Agent Orange. Contemporary Southeast Asia 29(1): 172–195.

41.

Palmer

Michael

Groce

Nora

Mont

Daniel

Nguyen

Oanh Hong

Mitra

Sophie

(2015) The Economic Lives of People with Disabilities in Vietnam. PLoS ONE 10(7): e0133623. https://doi.org/10.1371/journal.pone.0133623.

42.

Pham

Tai The

Nishijo

Muneko

Nguyen

Anh Nguyet Thi

Maruzeni

Shoko

Nakagawa

Hideaki

Hoang

Luong Van

Tran

Anh Hai

Honda

Ryumon

Kido

Teruhiko

Nishijo

Hisao

(2013) Dioxin exposure in breast milk and infant neurodevelopment in Vietnam. Occupational and Environmental Medicine 70(9): 656–662.

43.

Savitz

David A

Nguyen

Thang Minh

Swenson

Ingrid E

Stone

Erika M

(1993) Vietnamese infant and childhood mortality in relation to the Vietnam War. American Journal of Public Health 83(8): 1134–1138.

44.

Segall

Malcolm

Tipping

Lucas

Henry

Dung

Tam

Vinh

Huong

(2002) Economic transition should come with a health warning: The case of Vietnam. Journal of Epidemiology Community Health 56(7): 497–505.

45.

Singhal

Saurabh

(2018) Early life shocks and mental health: The long-term effect of war in Vietnam. Journal of Development Economics. https://doi.org/10.1016/j.jdeveco.2018.06.002.

46.

Staiger

Douglas

Stock

James H

(1997) Instrumental variables regression with weak instruments. Econometrica 65(3): 557–586.

47.

Stellman

Jeanne Mager

Stellman

Steven D

Christian

Richard

Weber

Tracy

Tomasallo

Carrie

(2003) The extent and patterns of usage of Agent Orange and other herbicides in Vietnam. Nature 422: 681–687.

48.

Strauss

John

Thomas

Duncan

(2008) Health over the life course. In: Paul Schultz

Strauss

John

(eds) Handbook of Development Economics, vol. 4. Amsterdam: Elsevier/North Holland, 3374–3465.

49.

Teerawichitchainan

Bussarawan

Korinek

Kim

(2012) The long-term impact of war on health and wellbeing in Northern Vietnam: Some glimpes from a recent survey. Social Science and Medicine 74(12): 1995–2004.

50.

Thayer

Thomas C

(1975) A Systems Analysis View of the Vietnam War: 1965–1972. Fort Belvoir, VA: Defense Technical Information Centre.

51.

Tirman

John

(2011) The Deaths of Others: The Fate of Civilians in America’s Wars. Oxford: Oxford University Press.

52.

Tran

Phung Kim

Viet

Husum

Hans

(2012) The legacy of war: An epidemiological study of cluster weapon and land mine accidents in Quang Tri province, Vietnam. Southeast Asian Journal of Tropical Medicine and Public Health 43(4): 1035–1041.

53.

Tuoitre (2018) Gần 19% diện tích cả nước ô nhiễm bom mìn, vật nổ [Near 19% of land in Vietnam is polluted by bombs and mines]. Tuoitre Online.

54.

Ugalde

Antonio

Richards

Patricia L

Zwi

Anthony

(1999) Health consequences of war and political violence. In: Kurtz

Lester

(ed.) Encyclopedia of Violence, Peace and Conflict, vol. 2. San Diego, CA: Academic, 103–121.

55.

VOV (2016) Over 100,000 people wounded by bombs and mines left behind after the war. Voice of Vietnam 31 March.

56.

Washington Group on Disability Statistics (2017) The Washington Group Short Set on Functioning (WG-SS) (http://www.washingtongroup-disability.com/wp-content/uploads/2016/12/WG-Document-2-The-Washington-Group-Short-Set-on-Functioning.pdf).

57.

World Bank (2011) World Development Report 2011: Conflict, Security and Development. Washington DC: World Bank.

58.

World Health Organization (2001) International Classification of Functioning, Disability and Health (ICF). Geneva: World Health Organization.

59.

World Health Organization & World Bank (2011) World Disability Report. Geneva: World Health Organization.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.07 MB