Virtual Water: from an explicative concept to a ‘prescriptive tool’


At the sixth World Water Forum in Marseille in March 2012, one target[1] over the twelve priorities for action was entirely devoted to the concept of virtual water (VW) and its corollary the water footprint, reminder of the increasing importance that both concepts have acquired in the recent years.

The concept of VW was introduced by Pr. Tony Allan in the 1990s as “the water embedded in key water-intensive commodities such as wheat” (Allan, 1997), in other words “the water needed to produce agricultural commodities” (Allan, 2003: 5). The underlying idea is that water-short economies – such as the ones of the Middle East that Allan studied extensively – are able to face water scarcity by importing water-intensive goods (i.e. VW), hence avoiding mobilizing their own scarce freshwater resources to produce these commodities themselves (Allan, 1997; 2002).

A derived and closely related concept is the one of water footprint: “The water footprint of an individual, business or nation is defined as the total volume of freshwater that is used to produce the goods and services consumed by the individual, business or nation” (Chapagain and Hoekstra, 2004, p. 11). To clarify the distinction between the two concepts, Roth and Warner (2007:9) suggested that while VW is related to the volumes of water used in production, the notion of water footprint rather describes the volumes of water necessary to support consumption.

Originally, the concept of VW was an explicative one, developed to account for the absence of water wars in the Middle East (Allan, 2002). The main point of Allan is indeed that water-scarce countries depending on other countries’ water resources can find a solution other than war or basin cooperation by importing water-intensive agricultural products. Indeed, importing a good amounts to import “virtually” the water needed to produce it. Allan (2003: 5) estimated that the MENA region imported in 2000 at least 50 million tons of grain annually, i.e. the annual volume of freshwater flowing into Egypt down the Nile.

Gradually, this analytical concept has evolved from its original explicative function to become a ‘prescriptive tool’ (Warner, 2003: 63). The main idea, developed by those advocating VW trade as a policy instrument, is that international VW trade, if optimized, could help saving water both at the national and at the international level. At the national level, it is recommended that water-short countries import water-intensive goods in order to overcome their water scarcity (Hoekstra, 2010: 7). Hoekstra (2010: 8) quotes Jordan as the perfect example of a country who managed to cover its water shortage by externalizing its water footprint. At the international level, VW trade is viewed as a way to save water globally by exporting water-intensive goods from water rich countries to water scarce countries (Chapagain, 2006[2]). Saving water globally is particularly essential in a context of “rapid and unchecked population growth” (Turton, 1999: 4) where there is certainty neither about the future global freshwater availability nor about the capacity to use water effectively at the global scale (Allan, 1997: 7). 

But, is this ‘conversion’ of a concept into a policy instrument a straight process, exempted from unwanted consequences? First of all, from an economic perspective, it is worth noticing that VW strategy may not be available for the poorest countries since they might not have the economic resources to import VW even if they desperately need it (Warner, 2003: 132; Chapagain, 2006: 122; Warner and Johnson 2007: 72). There are also political barriers to the implementation of VW trade strategies: governments may not want to implement strategic VW trade because of the issue of food security (Warner, 2003; Chapagain, 2006; Roth and Warner, 2007:10). “The political will to abandon the paradigm of self-sufficiency” is therefore a condition for strategic VW trade to be applied (Horlemann and Neubert, 2007; Hummel, 2005). Last but not least, ecological risks exist and should be taken into account. Regarding water degradation issues, it is crucial to note that VW trade will be beneficial only if this is blue water (groundwater and surface water) that is saved in water-scarce regions at the expensed of green water (evapotranspiration) in water-rich regions (Aldaya et al., 2010). If this is not the case, VW trade can have disastrous consequences as the example of the Aral Sea illustrates it dramatically: between 1960 and 2000, the Aral Sea lost approximately 80% of its volume because of the abstractions of water in the two rivers feeding it in order to grow cotton in the desert (Chapagain, 2006: 98). A more general environmental degradation may even occur with strategic VW trade given the fact that  exporting an agricultural good means exporting various factors of production besides water. In the case of Brazil for instance, the exportation of much more VW, even if perfectly possible in terms of water quantity, might come at the expense of the deforestation of the Amazon forest as many more hectares of land would be needed (Roch and Gendron, 2005). Partzsch and Schepelmann (2005, in Horlemann and Neubert, 2007) even contend that we should speak of “virtually traded land” along with VW trade.

Our point is not to dismiss VW as insignificant. VW is and remains an essential concept, both as an explicative instrument indispensable when assessing water resources globally and as a driver of policy awareness as it undeniably managed “to gain the attention of public officials and policy makers” (Wichelns, 2009: 2217). Yet, “policy value is found in complexity, not simplicity” (Wichelns, 2011: 643) and VW cannot be used alone as a policy tool, otherwise it could even become “dangerous” (Roch and Gendron: 2005: 282). Water, as an economic, political, social and ecological good can only be captured by a policy encompassing these different dimensions.

[1] Target 3.2: “Adjust pressure and footprint of human activities on water”

[2] See also Chapagain and Hoekstra, 2003; Hoekstra and Hung, 2002; Chapagain and Hoekstra, 2008.