Water productivity is broadly used as an indicator to measure the success of policies aiming at efficient water management. Focusing on the agricultural sector, this paper provides the first critical analysis of the appropriateness of water productivity for this type of assessment. We apply Logarithmic Mean Divisia Index (LMDI) decomposition analysis together with descriptive statistical analysis to test our main hypothesis that is natural resources do not generate value added, and therefore changes in water productivity are related to well-known drivers of economic growth rather than to improvements in water efficiency. We use three different models with different levels of decomposition detail to identify the drivers of water productivity changes. Results show that indeed there is a weak relationship between changes in water productivity and water efficiency. In contrast, changes in water productivity are driven by changes in labor productivity and capital intensity. Thus, we discourage policy makers from using the indicator to monitor the progress on efficient water management or on decoupling between economic growth and water consumption. Rather, indicators independent from economic development such as biomass per volume of water should be used and the absolute volumes of water appropriated by society monitored.

水生产率被广泛用作衡量旨在有效进行水管理的政策成功与否的指标。本文以农业部门为重点，对这种类型的评估对水生产率的适宜性进行了首次批判性分析。我们将对数平均除数指数（LMDI）分解分析与描述性统计分析一起使用，以检验我们的主要假设：自然资源不会产生增值，因此水生产率的变化与经济增长的知名驱动因素有关，而不是与提高用水效率。我们使用三种具有不同分解细节级别的模型来确定水生产率变化的驱动力。结果表明，水生产率和水效率的变化之间确实存在微弱的关系。相反，劳动生产率和资本强度的变化驱动着水生产率的变化。因此，我们不鼓励政策制定者使用该指标来监测有效的水资源管理或经济增长与用水之间的脱钩方面的进展。相反，应该使用独立于经济发展的指标，例如每体积水的生物量，并监测社会所使用的绝对水量。我们不鼓励政策制定者使用该指标来监控有效的水资源管理或经济增长与用水之间的脱钩情况。相反，应该使用独立于经济发展的指标，例如每体积水的生物量，并监测社会所使用的绝对水量。我们不鼓励政策制定者使用该指标来监控有效的水资源管理或经济增长与用水之间的脱钩情况。相反，应该使用独立于经济发展的指标，例如每体积水的生物量，并监测社会所使用的绝对水量。