19 February 2016

Four billion facing severe water scarcity? I think not.

Update (18 March): My letter (the post below) has been linked to over at Science Advances. Mekonnen and Hoekstra have "declined to respond."
This article has got a lot of attention. I think the attention is undeserved,* so I wrote this e-letter (an online comment on a published article)

Dear Editors and Readers,

Mekonnen and Hoekstra estimate scarcity based on physical models comparing water flows and population densities. From these models, they conclude that "four billion people [are] facing severe water scarcity." This title has generated headlines in the media, but it is misleading to the public. Indeed, it is even misleading to readers of the paper because management and governance -- only mentioned in passing -- are important, and perhaps determinant, factors in converting physical conditions into actual risk of shortage.**

Singapore and Israel, for example, have some of the lowest levels of total renewable water resources (108 and 227 m^3/ per capita per year, respectively), but the populations of these countries are not known for suffering from water risk. That is because their governments have been extremely proactive in converting natural supplies into useful supplies.

In their 1994 book, Rules Games and Common Pool Resources, Elinor Ostrom, Roy Gardner, and James Walker note that a "common pool resource situation" can turn into a "common pool resource dilemma" if (a) current strategies are leading to suboptimal conditions and (b) institutionally feasible alternatives exist that can improve on those outcomes (pp 15-16). As a water economist, I interpret their framework to mean that a change in governance or management can remove the dilemma, i.e., reducing risk to an acceptable, non-harmful level. That's why I wrote Living with Water Scarcity (2014). I wanted to make it clear that poor physical conditions did not necessarily result in water risk.

Curious to know more? I ran a simple regression of "Access to an Improved Water Supply in Urban Areas" against Total Renewable Water Resources (per capita), Freshwater Withdrawal versus Total Renewable Resources (two measures similar to those used in the paper), World Bank data on Control of Corruption, Effective Governance, and Regulatory Quality, and GDP per capita. You can guess that the latter 4 variables control for governance and wealth. Access to an improved supply may be a flawed measure of actual risk of water shortage, but it seems to be the closest variable we can find to a problem like "facing severe water scarcity," so I used it.

In a simple regression of Access against the two water availability variables, both were significant but only explained 1 percent of the variation (R^2 = 0.01). When I added the governance and income variables, R^2 jumped to 0.34 (the variables, as a group, explaining 34 percent of of the data variation in "Access to an improved supply"). More importantly, the physical variables dropped into insignificance, and variables for income and effective governance were quite significant. This hasty regression is not the final word, but it should be an obvious hint to the importance of governance and income on water supplies people care about -- and a hint that physical water conditions have little impact on those outcomes. (Data and regression results available at kysq.org/docs/WaterGov.xlsx.)

Four billion people are not facing severe water scarcity any more than seven billion people are facing severe food scarcity. In both cases, the difference between initial and final conditions is determined by institutional competence, i.e., good governance and management. These points are made by authors cited by Mekonnen and Hoekstra. Rijsberman (citation 6) says, "water will be a major constraint for agriculture in coming decades and particularly in Asia and Africa this will require major institutional adjustments." Wolfe and Brooks (citation 7) say "perceiving scarcity mainly in physical terms limits opportunities for policy-making and approaches for capacity building." I would have liked to see more of these perspectives in the main article, which asserted "four billion people facing severe water scarcity" without very good evidence. Indeed, I cannot even type "physical water availability is a necessary condition for scarcity," as Cherrapunji -- famous for being "the world's wettest place" regularly suffers from water shortages. Why? Poor water management.

Mekonnen and Hoekstra conclude by advising that "proper water scarcity assessment, at the necessary detail, will facilitate governments, companies, and investors to develop adequate response strategies." This advice is followed by suggestions of raising agricultural productivity and measuring water footprints. Although these recommendations make some sense, the first is not known to reduce risk of shortage (saved water is also used), and the second seems to reflect the authors' affiliation with the Water Footprint Network more than other, arguably more important responses, e.g., limiting water use in basins, increasing food imports to stressed basins, and -- above all -- improving water governance. These first two responses are mentioned in the article but the last is not. I am writing with the hope that this option will receive more attention.

David Zetland, PhD
Assistant Professor of Economics
Leiden University College
Den Haag, The Netherlands

* For older posts on (often useless) footprinting, click here.
 ** Michael Campana calls attention to the paper's weak treatment of groundwater

Addendum: This article on bullshit in science strikes a chord with me on this topic.

Addendum (12 March): Here's a great comment [pdf] from an ecological economist on the missing elements in footprinting discussions


  1. It would be very interesting to use subnational data to see which areas governments prioritize on with regards to providing them with access to improved water supplies. See whether selectorate theory can predict which areas are supplied, and which areas are not. (While stating the obvious), quality of governance is of course a political outcome, so it would be interesting to investigate the political side to water governance. Do you know about studies or datasets that have looked into this?

  2. Fernando de Barros22 February, 2016 15:29


    Where did you find the data on the renewable water resources availability in Israel and Singapore? The numbers are different from the list I have (World Bank - wdi.worldbank.org/table/3.5).


    1. Fernando,

      I used the figures in the referenced XLS, and the link to the data source (FAO, Aquastat) is there...

      Curious if the sources agree, or at least correlate :-\

  3. David, you're right on with the point that only good governance can help a community or nation manage water well. That being said, I think you missed a couple of key points here. One is the fact that this paper is not looking at water availability alone; instead, it is the proportion of the available water that is being consumptively used. We have found in our own research that there is a strong correlation between water "depletion" (ratio of consumptive use to availability) and the frequency and intensity of water shortages. In that light, these authors are at least focused on the biggest driver of water scarcity and shortages (note that water footprint is equivalent to consumptive use). However, the huge problem with the Mekonnen and Hoekstra analysis is their assumption that 80% of the available water is reserved for environmental purposes. While that's an admirable aspiration (and one that I myself have championed), to say that people are experiencing "scarcity" when there's 80% of the renewable water supply still flowing downriver or into the aquifer is just plain ludicrous. I can't believe that the national media and even Circle of Blue didn't catch this misleading assumption! For a much more accurate reading of water depletion, see our very recent paper in Elementa: https://www.elementascience.org/articles/83

    1. Brian -- thanks for the additional insight. I agree, of course, that use needs to be matched to supply (supply, demand!) to put shortages in perspective, but the role of storage, recycling (new supply!) and other management techniques did not get proper attention (esp with their bottom line).

      The role of environmental flows (as a source of demand against gross supply) is indeed enormous (since the beginning of human times), and I'm quite surprised that they imposed such an artificial (unrealistic) constraint.

  4. These are all good and valuable insights. Another basic mistake this study seems to make is to consider water scarcity only by looking at the water that flows in the rivers (modeled). They seem to ignore the part of the water balance where rain falls and feeds natural vegetation and soils. They miss both the rainfall and evapo-transpiration components of the water balance and only seem to consider river flows as input for products and their water footprints. Automatically this means that they have less water circulating in their footprints, and that their modeled scarcity might be much higher than in reality. Or did I miss something?

  5. María J. Beltrán03 April, 2016 18:04

    I think that much effort has been devoted to developing the water footprint (WF) methodology but very little to critically examining the implications of the use of virtual water (VW) and water footprint estimations as policy tools, and little (or no) attention has been devoted to analysing both virtual water and water footprint concepts as constructions supported by pre-existing ideas or ideologies, rather than a part of an independent objective reality.
    David I share you criticism about how the prescription resulted from the WF assessment undertook by Mekonnen and Hoekstra conclude that “four billion people [are] facing severe water scarcity” ignores the role of water governance.
    However, in your criticism you do not consider the current discussions on the social construction of scarcity. I argue that when the prescriptions resulting from VW and WF estimations conclude that a region suffers from 'water scarcity', it is crucial to examine the concept of scarcity that has been used in these analyses. Depending on which concept of scarcity is used, different economic (and water) policies will be suggested, which have different social and environmental consequences. Yet it is important to remember that 'scarcity' has often been used to increase the expansion of the water supply. Thus without the examination of the power relations and social constructions that lie at the heart of VW and WF analyses, these studies might promote a discourse that does not challenge the hegemonic construction of water scarcity in nature.
    In other words, the WF studies, depoliticising the idea of scarcity, reproduce a view of the world that obscures power relationships and other values unrelated to water scarcity.
    In short, I think is important to address what are the power relations and social constructions that shape VW and WF concepts.

    1. @Maria -- totally agree. Thanks.

      From Living with Water Scarcity:

      "Scarcity is a perception. Shortage is a fact. Most of us deal with scarcity every day. We spend our time going places, doing things and seeing people. We spend our money on products and services. We wouldn’t mind a bit more time and money, but at least we get some of what we want. Shortage is worse than scarcity because you can’t get any of what you want, even if you have time or money."


Spam will be deleted. Comments on older posts must be approved.
If you're having problems posting, email your comment to me