Page 103 - Climate Change and Food Systems

Bush, 1987; Vaux and Howitt, 1984; Medellin- Azuara et al., 2008; Olmstead, 2010a). Still, modelling water markets is not easy, as water
is not a typical commodity but a resource
whose use is geographically bound and whose management is intertwined with strong institutional arrangements. As a result, water prices are generally imperfect signals of resource scarcity
and rarely equate marginal water values across users (Olmstead, 2010b). Water prices are typically administratively determined, are unevenly applied, are often based on annual rather than usage based fees, and rarely take full economic value into account (Saleth et al., 2012).11
A number of institutional factors affect the development of water markets, including property right arrangements (often tied to historical and current entitlements), and legal and physical
limits on water transferability (Saleth et al., 2012). Externalities can also be an obstacle to water pricing such as the problem of “return flows” where irrigation water not lost to evapotranspiration either recharges groundwater aquifers or augments surface water flows within a basin, posing difficulties for pricing water usage (Libecap, 2011). Transaction costs (physical infrastructure, search, legal) are other important barriers to trade in water markets (Olmstead, 2013).
A big challenge in modelling water under climate change is the lack of water price datasets with sufficient geographic scope and which
hinder good estimates of the price elasticity of water demand (Olmstead, 2013). Thus, much of the economics literature on water demand has focused on the econometric estimation of demand parameters, including price elasticity.12 Estimating
11 Even in the US studies show that aridity and marginal price levels are negatively correlated (Olmstead, 2013).
12 A meta-analysis of 24 US agricultural water demand studies performed between 1963 and 2004 suggests a mean price elasticity of -0.48 (Scheierling et al., 2006) with high variability in reported values.
A review of studies conducted in developing countries suggests that residential price elasticity is in the range of -0.3 to -0.6, similar to the range estimated for industrialized countries (Nauges and Whittington, 2010).
water demand curve remains difficult owing to water-pricing practices. Farmers who withdraw water directly from surface sources usually incur an energy cost to convey water for irrigation, but do not typically pay a volumetric charge for the water itself. Many agricultural water demand curves are estimated for groundwater, using energy costs for pumping to construct a water price variable. While the economics literature contains many estimates of agricultural water demand elasticity, the available data are rarely of sufficient quality to estimate demand functions (Olmstead, 2013).
Water modelling at the river basin level can employ either optimization or simulation type models. Optimization models are used more often because they require less data than simulation models and can avoid the challenges of enumerating individual sets of feasible, optional rules. Moreover, optimization models can automatically implement autonomous adaptation to external shocks (Bell et al., 2014). On the
other hand, simulation models may evaluate specific adaptations (such as reservoir schedules, infrastructure improvements or policy choices) across different scenarios or even as discrete choices within the simulation.
The few available estimates of the economic impacts of the water resource effects of climate change make one of two assumptions about adaptation: no adaptation will take place; or water markets will respond in a dynamically efficient manner, maximizing the net benefits of water resources over time (Olmstead, 2013). Both sets of assumptions are problematic since water is rarely managed purely through markets. Examples of models that incorporate water include the IMPACT- Water (Rosegrant et al., 2002), which is used to assess adaptation options such as increased water storage and improvement in water use efficiency
at the basin and global levels (Nelson et al., 2010; Zhu and Ringler, 2012). The second is WaterGAP (Alcamo et al., 2003) which explicitly models water allowing for multiple water sources, sectoral demands, water infrastructure capacity and water policies.
Global water use modelling (GWUM) is fairly new and is much less advanced than global
chapter 3: economic modelling of climate impacts and adaptation in agriculture: a survey of methods, results and gaps
83