High levels of poverty, preventable disease and conflict are the prime causes of premature male death in developing countries – (Lee (2004:1) and Chant (1997:92)) make this point as well – creating the greatest number of widows least able to cope with the material consequences of widowhood. The high incidence of premature male death in developing countries is therefore a major personal economic and social status concern for women of all ages, especially for younger women and married girls. For developed countries, where most widows are retirement age or elderly, the main cause of widowhood is the differing life expectancy between men and women. Large numbers of elderly widows in developed countries are a significant concern since many have depended financially on their husbands and lack sufficient provision for the loss. The most significant issues for this group are health and the costs associated with health and care provision.

Global death patterns

Premature deaths are defined as those in the age range 15-59 years. With respect to widows, the younger end of this age range is valid given the higher rate of teenage marriage in many countries. Statistically, premature deaths in men are of direct interest to any study of widowhood focusing on causes and consequences, because, as indicated by the widows’ incidence model (below), the age difference in life expectancy between men and women significantly affects the number of widows in a population at any one time. Qualitatively, the age range in which women become widows is of paramount interest for the formulation and targeting of social protection policies and concrete assistance, because the needs of widows vary dramatically with age. Age can be analysed in terms of four status categories:

  • child widows without their own children,
  • widows with young children under the age of 18,
  • widows who are not yet elderly (especially those in the reproductive age range below age 50) with adult children,
  • elderly widows, including those who are no longer economically productive.

The only available comprehensive data that offers a clear global picture of mortality trends by sex and age is shown in Table 4.1: this shows the percentage of deaths for males and females in the age range 15-59. The Global Burden of Disease study (ibid.) reveals that ‘low- and middle-income countries account for a comparatively large number of deaths at young and middle adult ages: 30 percent of all deaths occur at ages 15 to 59, compared with 15 percent in high-income countries.’ This difference is caused by the greater incidence in developing countries of preventable diseases, defined as ‘communicable diseases, maternal and perinatal conditions, and nutritional deficiencies,’ which the GBD study termed ‘Group-I’ causes of death. Nearly all such deaths are in low- and middle-income countries as revealed in Table 4.1. In addition, preventable non-communicable diseases – for example, those caused by parasites in inadequate housing conditions – compound the effects of malnutrition and are thus much more likely to kill.

Globally, the total number of premature deaths in 2001 was 56 million. Of these, 10.5 million or 18.8 percent were deaths of children under five years of age, with four million mortalities below the age of one month. Almost all (99 percent) of these combined child deaths occurred in low- and middle-income countries (ibid.)
The 2010 Study estimated the number of under-five deaths of widows’ children to be around 1.5 million globally (Loomba Foundation, 2011).


1 Poverty

Table 4.3 shows the 2015 estimates for widows in extreme poverty by regions and globally. The adjusted 2010 data on extreme poverty is provided for comparison. In 2015 only extreme poverty estimates are presented as moderate poverty data is not available. The data shows that in 2015, an estimated 38,261,345 widows are living in extreme poverty where basic needs are not met. This is likely to be a minimum figure. The adjusted 2010 estimates show that 49,205,641 widows lived in extreme poverty at that time. This is a 22.2 percent decrease over the period, which mirrors the global drop in all persons in extreme poverty of approximately 20 percent in the World Bank’s 2015 forecast (World Development Indicators 2014).

It should be noted that due to the composite nature of the data used to estimate extreme poverty by region, these estimates must be treated with caution. Qualitative data suggests that in many countries widows typically experience worse poverty than other groups who are in the same economic bracket, so their situation is unlikely to be fully reflected in the whole country population data on which the poverty estimates are based and thus the decrease in numbers may be overestimated.

With that proviso, the region with the greatest change in extreme poverty figures over the period is East Asia and Pacific with a 58.6 percent decrease; Central Asia has seen a 51.5 percent decrease; Central and South America and Caribbean 50.4 percent, with South Asia and Sub-Saharan Africa reporting smaller drops at 12.1 percent and 11.0 percent. In Europe and Russia, the number of widows in extreme poverty rose by 47.2 percent, North America 37.4 percent, and Middle East and North Africa 6.2 percent. Globally, the percentage of widows in extreme poverty dropped by 22.2 percent. The global total of 38,261,345 widows in extreme poverty represents 14.8 percent of the global total of 258,481,056. Given the limitations of the data, these estimates may have understated the true percentage of extreme poverty in nearly all regions (see Appendix 3 for details on the data used). The groupings used in the Table are those employed in the World Bank source data.

Poverty acts as a primary cause of widowhood in three ways:

  • through hazardous work (often the only option for poor males),
  • personal behaviour or psycho-social tendencies causing early death, and
  • poor health and disease caused by inability to afford required nutrition and medical care in countries without effective free healthcare, lack of adequate housing, and lack of clean water and sewage systems.

Poverty here is defined as a range of deprivations, not simply lack of income, similar to the ‘capabilities approach’ created by development economist Amartya Sen (Sen (1999), pp. 72-76, 87-110, 131-138). Sen points out that $1 million dollars in cash is useless for an illiterate rural woman in a developing country. His idea is that there is more to determining well-being than money. While this approach is used by the United Nations in the annual Human Development Report series and elsewhere, income as a measure of poverty is still the most commonly used measure, probably due to its mass intuitive appeal and ease of calculation. It is currently a standard tool of poverty measurement in developed as well as in developing countries in the form of the income-measured poverty line (Laderchi, Saith and Stewart 2006: 19). It has also gained new visibility through the World Bank’s regularly reported ‘US $1 and $2 a day’ poverty measures, which are focused primarily on countries outside Europe and the OECD. These measures however have clear limitations as explained below and the resulting estimates of people in poverty should therefore be treated as minima.

The World Bank measure uses absolute poverty, which it also calls extreme poverty, and moderate poverty, which it calls “a higher poverty line used as a proximate vulnerability threshold to identify households who are not suffering absolute material deprivation, but are vulnerable to poverty. Although it seems somewhat arbitrary, it does bear some relation to empirically observed vulnerability to poverty” (World Bank 2005: 50).

This is contrasted with the relative poverty measure used by most developed country governments: ‘relative poverty defines poverty in terms of its relation to the standards which exist elsewhere in society…’,(Spicker, Alvarez and Gordon 2007: 169), with the relative poverty line measured as 50 percent of median income. A median income in one country, though, is not the same as in another, hence its definition as relative poverty.

The research department of the OECD Secretariat compiles absolute poverty statistics of OECD countries, which are used here based on an OECD study by Forster et al (2005) (), though this differs from the World Bank measure () as a statistical definition. Yet, there is an indication that the OECD measure is more or less a classic absolute poverty measure, when it states that ‘basic needs [do not include] child-care costs, for example…’ (Forster and d’Ercole, 2005: 21).

Absolute or extreme poverty looks at what is the minimum required to maintain physical well-being. This covers nutrition, shelter (together with heating where regionally appropriate), and clothing. ‘Moderate poverty generally refers to conditions of life in which basic needs are met, but just barely’ (Sachs 2005: 20). Thus, expenditure of between $1 and $2 a day could include very basic household items required for cooking, very basic healthcare and childcare. The concept of extreme poverty refers to inability to maintain core well-being or subsistence and stable good health. This therefore considers what resources are required, depending on regional conditions (e.g. climate), to attain the same state of core well-being whether one lives in Siberia, New York or Papua New Guinea.

Regionally this means that, according to the World Bank’s poverty measurement methodology, extreme poverty and moderate poverty are measured as living below US $1 or $2 a day respectively in the regions of South Asia and Sub-Saharan Africa, parts of East Asia, the Pacific, and South and Central America. For Central Asia, Eastern Europe, Russia, parts of South and Central America and East Asia – regions which suffer from extreme cold and so require additional expenditure in items such as warm clothing – under $2.15 a day constitutes extreme poverty, and $4.30 measures moderate poverty (World Bank 2008b and Alam et al 2005).

While young and elderly women can fall into poverty on the loss of a husband or partner in developed countries, in developing countries the consequences are potentially far more serious because of unfavourable economic and social contexts. Therefore, the longer a husband’s death is delayed, the greater potential there is to accumulate assets to cushion future household economic emergencies.

Families in developing countries – more so than low-income families in most developed countries, where the state often provides social protection measures such as healthcare free at the point of delivery, unemployment benefit, tax relief and housing subsidies – must in the absence of state support rely on building up a stock of assets to cushion against income-depleting events (see the literature on ‘livelihoods’, e.g. Ellis (2000); Ellis and Freeman (2004); Whitehead and Kabeer (2001)). This adds to the need to save for culturally mandatory high expenditure events such as weddings and funerals in many societies. If a husband dies prematurely, asset-stocking ability is severely affected, particularly where low wages are the norm for unskilled women and lower than men’s for the same job, or where female wage employment is socially restricted or not permitted at all. For those in extreme poverty it is impossible to save since daily income is insufficient to meet basic needs and government provision is typically inadequate or non-existent.

While attention is focused primarily on developing countries given their extreme consequences for poverty, some ostensibly developed countries are also at risk of this type of poverty. The USA and parts of the UK and Ireland have historically had a greater incidence of poverty and this remains arguably the case. They have been joined by the former communist states of Eastern Europe, Central Asia and Russia, where previous gains in living standards through universal social welfare provision have now been significantly eroded.

Figure 4.2 shows the relationship between extreme (absolute) poverty and the incidence of widows per country. The steadily increasing line is a ranking, in order of severity per country, of extreme poverty. The line that moves erratically up and down between zero and 20 percent – made up of connected individual country observations – is the percentage of widows per country corresponding to the rank of the country in extreme poverty. The numbered horizontal axis shows the cumulative number of countries in the world, totalling 193.

Poverty as a cause of widowhood is of special interest because of the prematurity of the widowhood caused. Premature widowhood is a particular concern because it has intergenerational consequences, characterised by women with dependent children. It is also a concern given the high reported incidence of widows under age 60 in developing countries. These women have been partly or wholly dependent on husbands and partners as the primary source of economic support and premature widowhood is more frequent in countries where state welfare in the form of income support and free healthcare is poor or not available at all, and in several regions, women’s paid employment is either low paid, or socially unacceptable. These are also regions where being a single woman is not socially acceptable. Becoming a widow under these conditions can therefore lead to a fall into poverty, or a deepening of existing poverty. Poverty therefore sets the scene and enables the conditions for people with low human and social capital and low economic and financial assets to descend into a range of deprivations. Poverty is a breeding ground for disease, and thus reinforces itself in a vicious cycle of deprivation (Hotez 2010) that makes it impossible for those affected to pull themselves out of the quicksand.

In turn, poverty or a fall in living standards, can lead to self-destructive psycho-social conditions, especially in men, leading to increased premature male mortality and hence early widowhood. Widows’ circumstances in most developing countries put them at high risk of poverty and additional acute direct threats to their well-being. Vulnerability also appears to be significant in East Europe, Central Asia, parts of Latin America and East Asia. While the picture is more favourable in most OECD countries, some of these have pockets of extreme poverty in which widows figure prominently.

To get a clear sense of what extreme poverty at the ‘$1 a day’ measure really means for people living below it, we can consider it in relation to the level of consumption assumed in the official poverty line used by the government of India. The comparison described below was made by the World Bank in a 2002 study, when its own extreme poverty line was estimated at £1.08. Available 1993-94 data on Indian consumption in money terms placed the Indian poverty line (measuring absolute poverty) at $0.22 – twenty-two US cents, which translated into daily food consumption of:

  • “Three scant plates of cooked rice, or 8-10 chapattis [flat bread like tortilla bread]
  • A half cup of cooked pulses [lentils]
  • A spoon of edible oil
  • A spoon of dried chilli
  • One medium-sized potato or onion
  • One cup of tea
  • A handful of brinjal [aubergine]
  • One half cup of milk
  • One banana three times each month
  • An egg every five days”

While relative prices have changed, and hence the composition of food consumption must have changed in recent years, the change is likely to be marginal; the 1993-94 food basket is still relevant for the poor today. After buying food, two additional rupees each day (about $0.06 cents) would be left over for items like medicines, school books, fuel for cooking, clothing, soap, durable goods, etc. And one-third of India’s rural population cannot even afford this frugal bundle (World Bank 2002: 13-14). Leaving the monetary indicators to one side, the items and quantities listed present a stark picture of what extreme poverty means.

The most recent available data for the Indian government’s and World Bank’s $1 a day poverty lines shows the percentage in poverty (below the poverty line) measured by the Indian government for year 1999-2000 at 28.6 percent, while the World Bank figure for India for 2004-05 is 41.6 percent. The Indian measure understates the extent of extreme poverty when compared to the World Bank’s measure.

To see what this ultimately means as a gauge of extreme poverty, as well as its scale and scope, a March 2008 Médecins Sans Frontières (MSF) press release stated that India is experiencing,

‘a major humanitarian crisis… South Asia is one of the world’s malnutrition ‘hot spots’, and in particular India carries the largest burden of illness in the region. One half of India’s children under five are underweight, while every day, six thousand children are lost to complications resulting from malnutrition and as many as 83 percent of women are anaemic.’

For Eastern Europe, Central Asia and Russia, which make up the former Communist region of the Warsaw Pact and Soviet Union, the World Bank produced the following assessment of ‘What Would Someone in the Region Living on Two Dollars a Day Consume?’ (Alam et al 2005: 52 and World Bank 2008b).

‘Average food expenditure needed to meet basic caloric requirements with the cheapest products available on the market is around $1.18 a day at 2000 purchasing power parity. Interestingly, it is found to be in a relatively narrow range from the cheapest basket of $1.15 a day in Tajikistan to around $1.22 in Kazakhstan. National data show that such allowances cover only very meagre baskets (composed predominantly of wheat, beans, milk, and oil). A person living at the poverty line of $2.15 a day would have been able to spend about $1 a day toward other needs. Such needs in the region primarily consist of heating and lighting. The approximate monthly electricity needed to light an apartment with three bulbs and run basic appliances (for example a refrigerator) is 150 kilowatt-hours. At prevailing prices of around two to five cents per kilowatt-hour, when converted into PPP (PPP exchange rates are typically three to four times market levels), and adjusting for family size (three to four per household), this amounts to $0.07 to $0.17 [cents] per day. Heating would require significantly more. Eurelectric’s (2003) ‘typical consumer’ on average requires an additional 350 kilowatt-hours per month.’ (Alam et al, 2005: 52, citing Eurelectric, 2003)

While large-scale incidence of poverty is concentrated in certain parts of the world, poverty remains a potent issue that has still not been satisfactorily dealt with in what are regarded as advanced countries. In all regions, part of poverty’s persistence is due to neglected policy areas stemming ultimately from overlooked issues, of which widows form an important part.

2 Work, disease and poor health

‘Work kills more people than wars’

The International Labour Organization (ILO) estimates that each year there are 2.3 million deaths of workers, or approximately six thousand a day, in part the result of the annual 270 million workplace accidents. Work deaths are defined as deaths arising from accidents and work-related illness. The ILO states that ‘work kills more people than wars’. Globally this costs four percent of GDP, while the accident rate has been increasing in developing countries where an unknown number of work deaths go unreported because a large proportion take place in the informal sector of the economy (International Labor Organization 2004).

Poverty frequently contributes to high premature male mortality through hazardous work, and hence to widowhood. Hazardous work is often exploitative, because in many countries, most low skill workers have no bargaining power. As a result, hazardous work-related deaths become more likely when men have low skills and low or no education, or limited local opportunities to use more advanced skills. Hazardous work also occurs where safety standards are poor or non-existent and production techniques low tech. An example is manual agricultural work, where highly toxic pesticides –  often having been banned in developed countries (Jacobs and Dinham 2003; Stillwaggon 1998: 263) – are used without protective clothing.

Stillwagon states that ‘studies of occupational safety and health suggest an enormous toll in mortality and morbidity throughout the developing world, with very high costs. An Indian study of occupational hazards in the countryside found that injuries constituted 14 percent of all morbidity. Extrapolating from the data collected in the sample region, researchers estimated serious injuries related to agriculture to amount to five million per year in India, with 500,000 deaths’. In Latin America, according a study by Professor Chant, ‘accidents are on the increase’ according to morbidity and mortality data (Chant 2003: 106).

A regular source of reported work deaths is in mining. Chinese and Russian mining deaths in particular are frequently in international news headlines. A western journalist investigating Chinese mining accidents reports:

‘[Mining] accidents are so common in China that their plight and that of tens of thousands of other mining widows has become one of the most sensitive issues facing the communist government. More than 5,000 Chinese miners are killed each year, 75 percent of the global total, even though the country produces only a third of the world’s coal. Working under appalling safety conditions, they are sacrificed to fuel the factories that make the cheap goods snapped up by consumers in Britain and other wealthy nations.’ (Watts 2005)

The ILO explicitly backs the conclusion that worker deaths are disproportionately a poor man’s burden: ‘Industrialised countries must take part of the blame.’ In fact, ILO specialist Jukka Takala says,

‘one of the trends is that industrialised countries are exporting their hazards to developing countries. Labour there is not only cheaper but also significantly less protected. Dirty and difficult jobs are left to the South [developing countries]. That includes mining, of course. So while the mining diseases commonly known as pneumoconiosis, including silicosis, [dust caused illness] have virtually disappeared in the industrialised countries, they are still claiming fresh victims every day in the developing world’. (International Confederation of Free Trade Unions 2002)

The ILO reports that six million miners in Brazil, nearly two million in Colombia and more than two million in India are exposed to conditions that cause silicosis. In Latin America, 37 percent of all miners are ill with silicosis. A 2011 report from Madhya Pradesh observes: ‘Miner tuberculosis creates village of widows in rural India’.


Preventable diseases are defined as infectious or communicable diseases spread from human to human and non-infectious diseases caused by individual behavioural responses embedded in social norms – these are referred to as psycho-social response diseases. In both cases, public health and state social welfare measures are required to counter their spread.

Disease as a cause of widowhood does not operate on its own, in a social vacuum. Social norms leading to socially learned behaviours are crucial facilitators of disease as a killer of men and hence as a cause of widowhood. When fused with poverty, they produce a lethal combination. Malnutrition caused by poverty – especially extreme poverty – reduces the body’s natural defenses to diseases such as HIV/AIDS, with skin on critical surfaces of the body becoming fatally impaired in its ability to resist disease. Data on causes of male death bears this out. Table 2.8 presents data on the top five global causes of death in males aged 15-59 in low- and middle-income versus high-income countries (Lopez et al 2006: 126-130 and 168-172, data from Tables 3B.1 and 3B.8). The top five group accounts for 56.5 percent and 69.7 percent of deaths respectively for these two country groupings. Causes of death heavily influenced by behaviour and low standards of living are represented for both low- and middle-income and high-income countries, notably road accidents, self-inflicted injuries and HIV/AIDS.

As noted elsewhere, while we have seen some improvement in HIV prevalence rates since the 2010 Study, the 2014 UNAIDS Gap Report shows that Sub-Saharan African countries still facing particularly severe HIV emergencies are Botswana, with an adult 15-49 age group infection rate of 21.9 percent, Lesotho (22.9), Namibia (14.3), South Africa (19.1), Zimbabwe (15.0), Swaziland (27.4) and Zambia (12.5). In India, total HIV/AIDS infections among all ages in 2013 was estimated by UNAIDS as 2.1 million (Wilson and Claeson 2009: 14-17) compared to South Africa 6.3 million, Nigeria 3.2 million, Kenya 1.6 million and Malawi 1.0 million. No data is available for China. Among European Union (EU) countries with data, Estonia (1.3 percent), Italy (0.3), Spain (0.4), Switzerland (0.4) and the United Kingdom (0.3) have the highest rates of HIV in the 15-49 age range, however HIV is much more likely to affect widows in Sub-Saharan Africa than those in the EU. Of 35 million people infected with HIV worldwide, an estimated 19 million do not know they are HIV positive.

The impact of unprotected sex and the link with HIV/AIDS through prostitutes, in addition to other forms of frequent changing of sexual contacts, is a major contributory factor to premature death.

‘HIV claims the lives of countless prostitutes and sex slaves each year, and unprotected sex with prostitutes is the single largest contributing factor to India’s rapidly growing HIV problem. Assuming 1.5 million prostitutes in India, nine hundred thousand (60 percent) would be HIV positive. If each prostitute provided sex to ten men per day and the infection rate were one in five thousand (0.02 percent), then each day, one thousand eight hundred men in India would become infected with HIV as a result of sexual intercourse with a prostitute. For each 0.01 percent increase in the infection rate, an additional nine hundred men per day would be infected.’ (Kara 2009: 271)

The effects of this and of similarly uncontrolled sexual promiscuity outside the sex trade on widows and their children are felt most acutely in Sub-Saharan Africa. East Asia has also been identified as a prostitution-HIV/AIDS hotspot with spillover into the general population through infection of family members, specifically wives: Thailand has been a noted case for many years given its huge sex industry, with reports of an increasing number HIV/AIDS infected husbands from middle class backgrounds.

Meanwhile a reported upsurge in prostitution linked to changes in men’s behaviour towards sexual activity has been reported in Vietnam. This change in Vietnamese male behaviour has been linked to the expansion of sex tourism in Vietnam since the start of its economic reforms in 1986 (Barry 1996: 144-156). The Philippines is also a noted prostitution hotspot and therefore a breeding ground for HIV/AIDS. HIV/AIDS is the fourth leading cause of death in low and middle-income countries and the leading cause of death for everyone in the 15-59 age group in Sub-Saharan Africa (Lopez et al 2006: 69, 71).

Chronic poor health

Chronic poor health, brought on by inadequate nutrition, access to healthcare, housing and lack of clean water and sewage infrastructure, is the leading outcome of poverty – with greatest intensity but not exclusively in developing countries. Crucially, chronic poor health aggravates the progression of serious diseases such as HIV/AIDS and tuberculosis (TB), the effects of which can be seen in differences between high and low income countries in male mortality figures and causes in Tables 4.1 and 4.2. Less prominent in the media and the public campaigns of aid agencies are diseases that cause chronic ill health, which truly are the diseases of poverty and these include the Neglected Tropical Diseases. The global picture of chronic poor health and the diseases of poverty have recently been highlighted in Scientific American magazine (Hotez 2010, see also Hotez 2008).

The key causes of people’s susceptibility to diseases of poverty were shown to be poor housing (exposure to disease-carrying parasitic insects), lack of clean water and sewage systems (water and sanitation), and hygiene knowledge and behaviours (together referred to as WASH – water, sanitation and hygiene). Slum residents in the Mathare informal settlement in Nairobi still believe that diaorrhea is caused by ‘evil spirits’, even while human faeces flow in open, shallow, improvised drains.

Many of the diseases of poverty are easily treatable and preventable, but tragically are not treated because those affected live in countries with low quality or non-existent public health systems and are unable to pay for private medical treatment. Notable among these countries, as pointed out in Scientific American, is the United States, where poor quality housing also plays a significant role in facilitating preventable disease. Disease becomes chronic poor health when a threshold is passed beyond which they can no longer be treated.

The former editor of the magazine International Agriculture Development, John Madeley, observes that ‘millions of the rural poor are handicapped by malnutrition, and are vulnerable to chronic illness and injury due to unfavourable working and living environments, including unclean water and poor sanitation facilities’ (Madeley 2002: 108).

This is backed by new health measurement techniques that calculate the number of years lost over a person’s life to poor health (Lopez et al 2006). In addition, chronic poor health greatly amplifies the mortality potential of opportunistic infections, evidenced by the high number of infectious disease deaths in developing countries (Table 2.7).

3 Suicide

Psycho-social response deaths in men are a pertinent area for consideration. The term psycho-social response deaths as used here distinguishes these deaths as inflicted by individuals on themselves, so focusing on the individual as the cause of death. The underlying cause of death, however, is from the male psychological response to reduced economic circumstances. This is a factor whether the deaths result from a prolonged process – as in the case of alcoholism – or when immediate, as in the case of suicide or non-military community violence.

A significant cause of premature deaths of males aged 15-59 is alcoholism as a psycho-social response to economic collapse, which has epidemiological characteristics due to its society-wide incidence. This has been observed in Russia since the early 1990s, following the collapse of the Soviet economy and welfare state. It has been so pronounced that the adult male death rate has increased dramatically, reducing male life expectancy to 58 years.

‘From 1991-94, the risk of premature adult (15-59) death increased by 50 percent for Russian males. It improved somewhat between 1994 and 1998, but subsequently increased’ (ibid.). The resultant increase in the numbers of widows can be seen in Table 4.5.

Thus it can be expected that there are many more younger age widows in Russia than is the developed country norm. Comparing Russian census data on widows for 1989, just before the collapse of the Soviet Union, and in 2002 after the transition of Russia to a market economy, an approximately three and half percentage point increase is revealed in the number of widows in the 15-49 age group.

Alcoholism is also frequently reported in India as a cause of death for men from low-income backgrounds, such as those living in slums or in communities where seasonally enforced breaks in employment – for example in agriculture, or during the monsoon for small scale fishermen – are the norm. On a 2009 visit to the state of Kerala, the president of India commented that ‘alcoholism is rampant’ where human development for women in health and literacy is comparable with developed countries (Hindu 2009: 1). It takes hold during off-season work periods, such as the monsoon, when a long period of bad weather prevents fishermen from going to sea.

With reference to Latin America, it has been observed that males are affected by ‘a general pattern whereby male morbidity and mortality is more likely than women’s to occur as a result of behavioural factors such as risk-taking, self-abuse, or attempts to abuse others, including drinking, smoking, car accidents and fighting’ (Chant and Craske 2003: 120, citing Mueller, Helena and Yunes 1993 and Zapata et al 1998) .

In many developing countries, widespread suicides have occurred among farmers as a result of deteriorating economic conditions in agriculture. India is often cited in discussions on rural economic distress in developing countries. The consequences for farmers’ widows – rarely focused on – are illustrated in the example below and in section 5.1.

‘Mr. Kelkar [a cotton farmer who hung himself out of despair over debt] had often talked farmers out of taking their lives in the state’s cotton growing belt of Vidarbha where, on average, one farmer commits suicide every eight hours. In other words, three women become widows here every day. Mr. Kelkar’s wife, Indira, is now one of them. She is left with the mammoth responsibility of paying off his debt while looking after their four children.’ (BBC 2006b)

In his book Stuffed and Starved, Raj Patel (2007) observes on farmer suicides across India:

‘Authoritative figures are difficult to come by at a national level, but the state of Andhra Pradesh, with a population of seventy-five million, has been recording rural suicide rates in the thousands per year. Nor is it a problem limited to Andhra Pradesh. The hinterland of Mumbai, where the city finds its food, has experienced a rocketing rate of farmer suicide. It’s a problem that has even hit India’s breadbasket. In Punjab, the epicentre of the country’s high tech agricultural ‘Green Revolution’, the United Nations scandalised the government when it announced that, in 1995-96, over a third of farmers faced ‘ruin and a crisis of existence…This phenomenon started during the second half of the 1980s and gathered momentum during the 1990s. It has been getting worse. According to the most recent figures, suicide rates in Punjab are soaring.’ (See also Mishra 2006, Panargya 2008: 152-154, Reddy and Mishtra 2008: 47-50, Shroff 2008)

Farmer suicides in India have been a feature for some time, and while they have received a great deal of attention in discussion by policy makers, researchers and the media (Patel 2007: 26, citing Phillips, Li and Zhang 2002), nothing suggests that decisive action to tackle it has been or will be taken in the near future.

Patel also cites evidence of suicides among agricultural workers in China, making up over fifty percent of the deaths in one sample, using a study of suicide data for the period 1995-99 (Patel 2007:25 and Murphy 2004:258). In the case of China, there is some evidence to suggest that these deaths may be linked to the break-up of the old collective agricultural system, where farm households in effect pooled resources, including heavy ploughing equipment provided by the government. Many rural Chinese widows are in turn unable to cope alone in agriculture. Patel finds evidence of higher rates of suicide among farmers and agricultural workers among countries as diverse as Sri Lanka, Australia, the United Kingdom and the United States.

Another global cause of death not covered in this report is drug addiction.

4 Road accidents

While not popularly associated with developing countries, Table 4.4 shows that road traffic deaths are in the top five causes of premature male death in both developing (low and middle income) and developed (high income) countries alike. Kevin Watkins, formerly head of research at Oxfam GB and former director of the United Nations Human Development Report, characterises the significance of road traffic deaths and injuries as a ‘global pandemic.’ He emphasises that ‘traffic injuries in developing countries kill on a scale of malaria or tuberculosis. For children between five and 14, traffic injuries are the biggest single source of death, and roads are second only to HIV/AIDS in killing people aged between 15 and 29… The world’s most dangerous roads are in Africa. Britain has a fatality rate of one death per 10,000 vehicles; in Ethiopia and Uganda it tops 190. Traffic deaths are climbing most rapidly in Asia and Latin America…’ (Watkins 2008). In another example, young men interviewed in Kenya explained they were not concerned with likelihood that their risky sexual behaviour might cause HIV/AIDS since they felt the probability of dying in a road accident was the same.

5 Conflict

Interstate war, civil war and political repression

Conflict as a cause of widowhood can be inter-communal, such as in Kenya in 1992, 1997, and 2007, due the contested presidential elections; in Northern Ireland since the 1970s; the case of the Gujarat massacre of Muslims in India in 2002; massacres and targeted killings in Indian Kashmir in addition to military confrontations; or the Shia-Sunni conflagration of post-Baathist Iraq following the US-led invasion. The death rate of young males is also aggravated by the high level of armed violence in Brazil, Colombia, northern Mexico and several US inner cities – all a byproduct of the international drugs trade. Other locations with high rates of armed violent crime are Chechnya, Kenya, Nigeria, Somalia, and South Africa.

More traditional forms of conflict, such as guerrilla insurgencies and military warfare, produce the highest levels of deaths, again impacting on widowhood. The Democratic Republic of Congo, Chechnya, Iran-Iraq, Afghanistan since 1979, Sri Lanka, the Lebanese civil war, and the US-led battlefield confrontations with Iraq in 1991 and 2003 followed by multi-party fighting, are all examples. The eight year Iran-Iraq War (1980-88) killed 500,000 Iraqi soldiers alone (Enloe 2010: 65). The effects of many years of war in Vietnam – including fighting with China after 1975 – continue to be felt through a large presence of widows in the Vietnamese population. The Eritrean war of independence that ended in 1991 caused an estimated 200,000 deaths (Smith 2003: 116-117, citing Tekle 1998: 1). Estimated total deaths (combatant and non-combatants) during the 1979-92 civil war in El Salvador were 80,000 of the total population of five million (Thompson and Eade (2007: 126). Guerrilla-style fighting, with the addition of terror tactics, has continued for the last 50 years between the Palestinians and Israelis, producing a steady stream of widows. More recently, Algeria has suffered an especially brutal civil war instigated by Islamic extremists that has claimed at least 150,000 lives between 1992 and 2005 (Liverani 2008: xxviii), in part through extreme forms of terrorism. The conflict in Indian Kashmir, with the presence of Pakistani, international and Kashmiri insurgents as well as the Indian army, continues to result in civilian deaths. Pakistan itself is current gripped by an intensified Taliban-inspired insurgency.

Table 4.6 documents the broad global coverage of war widows based on the prolific state of ongoing conflict, particularly warfare. The region most affected is Sub-Saharan Africa, with several ongoing civil wars, and major civil wars that concluded in the last ten years.

Since the start of the US-led invasion of Iraq in 2003 up to August 2007, a combination of combat deaths, unintended civilian casualties, widespread sectarian killings, illegal killings by security and military forces and violent crime, have produced an estimated 733,158 to 1,446,063 deaths according to one British polling organization.

On the battlefield, armies of developed countries can swiftly extract wounded soldiers by vehicle or helicopter and provide medical care in transit to a well-equipped medical base for immediate attention, and then by plane to a modern hospital for specialist care. Even so, the deaths of US soldiers during the Vietnam War amounted to 58,000, while post-9/11 operations have resulted in over 4,000 US soldiers killed to date.

In contrast, the armies of developing countries are often rudimentary. Battle and post-battle medical care of soldiers is often very poor, due to the lack of trained medical and planning staff, logistics limitations, and lack of specialist medical supplies. The war in the Democratic Republic of Congo in the 1990s, which involved the armies of several neighbouring countries, resulted in an estimated total number of deaths, male and female, of two million.

Combat deaths in developing countries are boosted by cheap but robust assault rifles and other cheap mass-availability weapons suited to low-technology warfare by semi- and illiterate soldiers. This translates into ill-coordinated battles at close range with fully automatic weapons and mortars – referred to as ‘poor man’s artillery’ because it consists of a tube, base plate, and bipod that can be easily stripped down and transported on a donkey or carried. Together, they maximise the potential for high casualties. This is perhaps best exemplified by the Angolan civil war from 1975 to 1994, when as many as 500,000 people died in the space of a two year period, 1992-94, towards the end of the war ‘through combat and war-induced starvation… in some of the heaviest fighting in the post-Cold War period… Although the government forces made some use of planes and tanks, most of this carnage was produced by light and medium weapons – mortars, light artillery, grenade launchers, machine guns, recoil-less rifles, assault rifles, landmines…’ (Klare 1997: 64-65).

A well-documented example that included large numbers of heavy weapons (tanks, armoured personnel carriers, heavy artillery, large missiles, aircraft, etc.) is the Iran-Iraq war of the 1980s. This saw forces confronting each other in World War One-style assaults, but with far deadlier weaponry, across open ground into concentrated enemy fire. The result was a vast number of war widows. Similarly, the Ethiopian civil war that concluded in 1991 involved huge amounts of these weapons, supplied among others by the then Soviet Union. Ethiopia government forces were described in that period as the largest in Sub-Saharan Africa (Klare 1997: 62).

Massacres and summary executions are also a source of conflict deaths. In Europe, when such events were thought to have been consigned to the Second World War era, the conflict in the former Yugoslavia reintroduced these types of atrocities. The most infamous example was at Srebrenica (Loyd 1999: 293-294 and Rohde 1997), where 8,000 men and boys were massacred by the Bosnian Serb army in 1995. In Beirut in 1982, when the Israel-Palestine Liberation Organisation (PLO) dispute spilled over into the Lebanese civil war, Christian militia fighters indiscriminately massacred Palestinian refugees in the camps of Sabra and Shatila. In the same year the Syrian army, under orders from President Hafez al-Assad, carried out a massacre to put down an uprising in the town of Hama, with fatality estimates ranging between 10,000 and 40,000.

The number of Iraqi civilian deaths from Saddam Hussein’s Baathist regime’s numerous internal crackdowns is unknown. The largest of these occurred after the 1991 Gulf War, focusing on the Shia and Kurdish communities in the south and north respectively. The previous decade, the regime had used poison gas against the Kurds, killing around 100,000 in the Anfal campaign. Similarly, the anti-insurgency campaign against the Mayans in Guatemala was described by the United Nations as ‘genocide’ (Steele and Goldenberg 2008: 11, Pilger 2007 citing the United Nations Truth Commission for Guatemala).

Internal crackdowns resulting in political ‘disappearances’ are another source of male conflict mortality and cause of widowhood. The best-known of these occurred in Argentina and Chile. El Salvador saw some of the worst political repression of regime opponents during its civil war from 1979-92, resulting in the disappearance of some 7,000 people. This has been described as:

‘…a particularly cruel civil war for the civilian population… In the cities the armed forces arrested, ‘disappeared’, tortured, and killed tens of thousands of people – professors, union organisers, health workers, slum dwellers, students, lawyers, and church workers. By 1984, the popular movement has been wiped from the streets; almost an entire generation of civil society leaders had been assassinated. In the countryside, the military undertook a scorched-earth policy to depopulate the zones in the north and east of the country held by the [rebels]. They razed homes, massacred entire communities, destroyed crops and livestock and carried out carpet bombing. By 1985 the [rebel-held] zones were largely depopulated, and one in five [twenty percent] Salvadorans was displaced within the country or had sought refuge abroad.’ (Thompson and Eade 2007: 126-127)

Open political repression, such as the 1989 Tiananmen Square massacre of students in China or the ongoing state of Zimbabwe under Robert Mugabe and the military government in Myanmar, continues to be a problem around the world.

One consequence of warfare is disease and starvation caused by disruption of food supplies and health services, resulting in deaths among the general population. This occurred with devastating effect in the multi-country war in the Democratic Republic of Congo which began in 1996 and extended internationally from 1997-2002, where civilians fled into the deep bush areas to avoid approaching armies, and are said to have perished in large numbers from lack of food and water (Legros and Brown, citing Rehn and Sirleaf 2003: 33). As a result, there are estimated to be 1.9 million widows in DR Congo in 2010, some as young as ten. Mass starvation also resulted from the civil war in Somalia in the early 1990s.

International Rescue Committee (IRC) research by a demographer Les Roberts estimates that there were 2.6 million conflict deaths from all violence and health-related causes (including  starvation) in the Congo between 1998 and 2001 (Prunier 2009). Roughly 86 percent of the deaths were estimated to have been from ‘disease and malnutrition.’ A newer estimate of Congo’s total conflict deaths (Coghlan et al, 2006) places the figure at 3.9 million for the period 1998-2004. In 2009 the same researchers published their latest survey of the country and found much higher death rates resulting from continued lack of basic health services. Mass starvation also resulted from the civil war in Somalia in the early 1990s when 300,000 died from lack of food and healthcare, and a repeat of this was threatened in 2009 (Burnett 2005, Black 2005).

While before-and-after time series data from the same source is unavailable for DR Congo, Table 4.5 shows the before-and-after effects of the Liberian civil war that began in 1989 and ended in 1997. While DHS data do not go beyond age 49, in the absence of other major calamities in Liberia during this period – including an absence of east African-levels of HIV/AIDS – it is evident that the high number of widows ten years after the end of the conflict in the 45-49 age group is a result of the civil war.

Genocide deaths

Genocide, the systematic and large-scale killing of non-combatants, is facilitated by war, and is sometimes its primary purpose. It is therefore the most serious war-related cause of death for the general population as, once organised and put into action, it is difficult to escape or stop. It is usually perpetrated in the context of a country-wide internal war. The most recent large-scale and well-known example is the 1994 Rwandan genocide, in which an estimated 800,000 to one million people were killed and this ‘also contributed to new wars that bedevilled central Africa into the twenty-first century’ (Shaw 2003: 211). As many as 50 percent of married women were estimated to have become widows, the majority of them over 30.

Table 4.6, also using DHS data, shows the effects of the genocide on the number of widows in Rwanda and the change in those numbers over time, ten years after the genocide. Notable is the extent to which the number of widows has quickly dissipated across all age ranges, to the point that the average for the whole age range 15-49, as a percentage of all women 15-49, has effectively reached its pre-genocide level. This is probably due to reduced rates of marriage and the generally very high mortality rate in the country between birth and age 30 (ORC-Macro International 2006, ch. 2).

Rwanda was not the first genocide in the region around that time. Before then, neighbouring Burundi experienced two genocides, according to the final report of the International Commission of Inquiry for Burundi discussed at the UN Security Council in 2002 – the first in 1972 and the second in 1993, a year before Rwanda. There had been waves of inter-ethnic massacres since the country gained independence in 1962. Between 1962 and 1993, one estimate suggests 250,000 people died.

In East Timor, formerly a part of Indonesia, 200,000 deaths have been reported for the period 1975-99, many as the result of massacres of non-combatants (Rosenberg 2003: 229). In Cambodia, the Khmer Rouge regime killed an estimated 1.5 million to three million people of a total population of seven million (ibid.). The Cambodian genocide has been described as ‘the most comprehensive of all modern mass killings, in the extent to which it touched all sections of the population within a given territory.’ At least one researcher refers to a higher than normal number of widows in Cambodia, at around 11 percent; an aid agency estimate made before 1994 was to have stated that approximately 80 percent of rural households were headed by women, most of them widows. The most recent census estimate for Cambodia places widows at 6.4 percent of the female population aged ten and above (1998 Census) (Shaw 2003: 166, Lee 2004: 3,  Chant 1997: 92, quoting O’Connell 1994: 68).

More recently, the actions of the Bosnian Serb army during the 1990s have been characterised by international officials as genocide, with total war deaths on all sides – most of them civilian and non-Serb – at over a quarter of a million (Shaw, ibid., 192).

In 2002, around 2,000 people were massacred in the Indian state of Gujarat, facilitated by the local police.

Deaths from landmines and other unexploded ordnance

A less well-publicised but equally important cause of male deaths occurs post-conflict, with serious economic implications for rural women and widows, as a result of landmines and other unexploded ordnance (UXO) deposited during combat, ranging from hand grenades to cluster munitions (often referred to as cluster bombs).

The most serious type of UXO is the anti-personnel landmine (Black 2005: 212). After landmines, the most insidious UXO is the mortar shell. Mortars are called poor man’s artiller’ because they are compact and relatively cheap; mortar shells are typically fired in large numbers. They can found in regular use in armies across developing countries. The main casualties from landmines and UXO are the rural population, typically farmers who disturb landmines and other UXO in the course of everyday work, often resulting in death.

The key characteristic of the UXO issue is its longevity, because clearance is slow and expensive, and because they typically remain active for many decades. For example, six hundred tons of UXO continues to be discovered and cleared every year from World War Two. Since World War Two, landmine technology has evolved so that many now include anti-tamper devices that cause them to explode if moved after they have been laid in the ground, so modern landmines must be destroyed where they lie. In addition, plastic or ‘minimum metal’ landmines have been developed to thwart traditional metal detectors, resulting in the use of sniffer animals, particularly rats and dogs. All this has made traditional dig-and-remove clearance impossible and contemporary landmine removal is thus very slow and expensive.

There are 33 countries, ‘where landmines constitute a major problem for the civilian population; the situation in five of those countries – Afghanistan, Angola, Cambodia, Iraq and Laos – must be categorised as an emergency on the basis of the scale of existing [post-war] casualties.’ Landmines deliverable from aircraft were deployed during the Soviet-Afghan war,  enabling a large increase in the numbers of landmines used. It is estimated that there are currently around 100 million landmines in the ground around the world.

In 2006 in South Lebanon, during the Israel-Hezbollah summer war, Israeli forces used cluster shells delivered by mobile heavy artillery guns on Hezbollah positions, carpeting these areas. Cluster munitions are large shells designed solely for killing people, not destruction of property; they contain a large number of small anti-personnel bombs inside. A British cluster rocket, for example, carries six-hundred ‘bomblets’, which are released from the cluster munition shell above ground in order to carpet a broad area. They have a high failure rate: many fail to detonate when first fired, only doing so later when disturbed. The combination of this high failure rate, small size (roughly the size of a large Coke can) and ability to cover a large area of ground, has caused them to be likened to landmines. Hezbollah also fired anti-armoured vehicle cluster munitions into Israel using Chinese-made Type 81 rocket launchers during the 2006 war.

This was only the latest episode in Lebanon’s UXO problem, as the case of a Lebanese widow’s battle with landmines makes clear. The UK-based Mines Advisory Group (MAG) – a leading UXO clearance NGO – documented the story of a Lebanese widow, now aged 60, who was hit by gunfire in a battle in her village during the civil war in 1977. In the same incident, her husband was killed outright. Her farmland was laced with UXO. She had nine children, the war pension was too little for the family to live on, so she started a small shop, which was looted several times during the fighting. In the early 1990s, using a knife, she started clearing her farmland of landmines. By 1997, she reached the limits of her property, but it was only finally made safe after MAG started work in 2007. (Story of Mrs. Em Saoud Mashmoushi, Bsaba village, Chouf Mountains, Lebanon) (Mines Advisory Group 2009).

The 2014 Gaza war and the ongoing civil wars in Syria, Iraq, Libya and Yemen will only increase the amount of UXO and Improvised Explosive Devices (IEDs) with a parallel rise in deaths.

Scale of conflict deaths

What is the scale and relevance of conflict as a cause of widowhood? Up to 1960, simultaneous conflicts globally averaged around 25. After this date we have seen a steady increase in conflicts, peaking at more than 50 in 1992, and falling to just below 30 in 2003 (Black 2005: 210, Fig. 13.1). In relation to the number of war widows, whereas it is possible to count the number of widows in censuses and surveys, data on the causes of widowhood is either not collected or, if collected, not made available.

Data on conflict deaths and injuries, including sex crimes, is extremely problematic. Data on rapes would be extremely valuable for estimating the scope and scale of health interventions, given the prevalence of HIV/AIDS in many conflict countries and the fact that many victims are women widowed as a result of the fighting. Conflict death and injury statistics are almost always very loose estimates in many developing countries, given the lack of wartime record-keeping systems. Employing professional demographers, as by IRC in the Congo example cited above, rarely happens. With respect to gender, ‘there are relatively little sex-disaggregated data on the impacts of armed conflict; and indeed the difficulty of collecting data in any conflict zone means that there is usually little dependable data at all’ (ibid., 211). Therefore, at the present time, there appears to be little or no data on the number of war widows with sexually transmitted diseases contracted as a result of military and random criminal actions.

Estimates for 2000 produced by Murray et al (2002) show significant excess male deaths, particularly in the 15-44 age range. This indicates the clear likelihood of a sizeable number of widows, since the high male death rate in is the prime male reproductive age group. One estimate suggests that the proportion of widows in conflict societies may be as high as 30 percent of a country’s total adult population (ibid., 213, Fig. 13.2, and Sorensen 1998).

There is no up-to-date global figure for the actual number of war widows, but some better-documented country-level examples stand out. These are Afghanistan with up to two million, Iraq 740,000 to eight million, and Rwanda 370,000 (Ministère des Finances et de la Planification Économique 2005).