Cooking the Numbers: The Misleading Use of Water Numbers

I'm a believer in the power of facts and numbers to help make public policy. I believe that without good science and without good numbers, it is difficult to make good public policy. And it is impossible to make good public policy with bad numbers.

The good news is that most policy makers understand this. The bad news is that there are people and groups (on both the left and the right) who work to produce or disseminate bad numbers as a way of influencing policy decisions in their favor. I've addressed these broad questions of the misuse of science in testimony to the U.S. Senate in 2007 and the Institute launched an Integrity of Science Initiative in 2005 to explore case studies of the abuse and misuse of science. We've described a number of remarkable abuses over the years.

Often, the generation and use of bad numbers comes down to what I describe as the "denominator problem." Numbers are often represented as a percentage or fraction or ratio of something: for example, a new "high efficiency" toilet uses around 20 percent (or 1.2 gallons divided by 6 gallons) of the water used by an old, pre-1994 toilet. Or the average American uses around 1500 gallons of water each day (calculated by dividing the fresh water used for everything nationwide by the total number of Americans).

But the choice of the denominator has a huge influence over the end result, and there are a growing number of examples where the choice is made for political purposes, to come up with a result more favorable to a particular political position. Here are two examples:

Water Number: 0.019% or something much larger.

What is the fraction of groundwater used for bottled water? A substantial amount of the bottled water sold in the United States — around 60 percent — comes from groundwater. This water is typically labeled "spring" water according to regulations set by the FDA. In recent years, there has been growing public opposition to the construction of large spring water bottling plants in small rural communities in Maine, Michigan, California, Colorado and elsewhere because of fear, and some direct physical evidence, that such large plants adversely affect local groundwater levels, flowing springs and local wetlands. In response, the bottled water industry, led by the International Bottled Water Association, launched a campaign (including testimony to state and federal legislators) arguing that there was no problem because "ground water withdrawals for bottled water production represent only 0.019 percent of the total fresh ground water withdrawals in the U.S." Ah, here rears the ugly head of the denominator problem. This number is probably very close to true. It is also completely irrelevant and misleading. The proper denominator should not be total U.S. groundwater withdrawals, it should be some measure of local groundwater availability, or use, or yield — a much smaller denominator. In this case, a bottled water withdrawal may be a very significant fraction of local groundwater. But by choosing a big denominator, the industry was attempting to disguise a problem.

Water Number: 80 or 41 percent?

Which number best describes the extent of Californian agricultural water use? A second example of a denominator problem is the number used to represent the use of water by the agricultural sector in California. There are two competing camps. The most common number, and the number I typically use, is that the agricultural sector uses around 80 percent of all of the water withdrawn for human use in California (34.2 million acre-feet per year agricultural use divided by a total human use of 43.1 maf). This is the typical approach used by water use analysts, internationally, nationally and locally. Indeed, the California number is the same as the global estimate of agricultural water use: 80% of the water humans use goes to agriculture. This number, to be polite, drives some in the agricultural community crazy because they think it makes them look bad. They insist on the use of a different denominator: instead of the total amount of water used by humans, they prefer the larger estimate of the total amount of water available in California. Using this number, agriculture only uses around 41% of the state's water — still a large amount, but much less dramatic than 80 percent. In this case (unlike the groundwater example above) both numbers are "right" — they just represent different things, including a different way of thinking about water. The agricultural industry prefers, of course, the larger denominator because it makes their use of water look smaller (40 percent instead of 80 percent).