Note that demand can be reduced in other ways, i.e., by requiring that technology be adopted (e.g., low flush toilets) or by a change in attitudes (e.g., people deciding that green lawns are not appropriate). Economists say that these reactions are shifts in the demand curve, while reactions to changes in price are movements along the demand curve. Shifts tend to last longer than movements, but both give the same result -- less water used.
These basic ideas may be too theoretical for some people, which is why I was thrilled that MU sent me an excellent paper demonstrating the concepts in action. In this 1997 paper [PDF], Loaiciga and Renehan discuss how Santa Barbara's implementation of radically higher prices during the 1987-1991 drought reduced demand.
The short story is this. When the drought started, water was priced at a flat rate, i.e., the price/unit did not change when people used more units. In 1990, after the drought began, the price schedule was changed to an increasing block rate (the first unit was $1.09; later units were $1.58, 1.97 and 3.01). After six months (as shortages grew worse), prices were radically steepened: first units at $1.09, then 3.27, 9.81 and 29.43/unit, as people used more. (A unit is 748 gallons or one-hundred cubic feet of water -- abbreviated HCF.)
Note that the median household in this period reduced its use until it paid a marginal price of $3.27/HCF, i.e., the price in the second tier that was higher than the lowest price by lower than quite expensive, third-tier prices.
The median user's behavior indicates (to me) that the very presence of increasing block rates is perhaps just as important (or more important) as the size of the jump between blocks, which were calibrated in a way that made them relevant -- and effective. (If the next block is too far away to affect usage, it's as if it doesn't exist at all.*)
Of course, prices were not the only means of affecting demand -- there were "conserve water" campaigns, watering restrictions, etc. -- but prices were a firm reminder to everyone (including those who didn't care for conservation ethics) that water was indeed in scarce supply. (Few people paid $29.43/unit, but they knew that water uses at that level were very "costly" to the community.)
What's interesting is the way that these prices permanently changed people's habits and attitudes. Before the drought, median household usage was 9 HCF/month (about 25.5 m^3/mo). During the drought, it dropped to 6.3 HCF/month.
Overall usage -- controlling for population -- dropped from 7.8 million HCF in pre-drought 1986 to a low of 3.9 million HCF in driest period (1991), i.e., by roughly fifty percent. After the drought ended, water use rose to 4.8 million HCF -- higher, but still only 62 percent of pre-drought levels.
The paper is packed with statistics, so here are a few more:
- The average price/HCF was $1.18 in 1986. It rose to $3.49 in 1991 and stayed at that level or higher.
- The City of Santa Barbara used this extra revenue to pay additional "drought-proofing" costs, e.g., building a desalination plant (now on "drought standby") and connecting to the State Water Project.
- Before the drought, 15% of households used more than 30HCF. During the drought, none did. After the drought, less than five percent of households returned to these consumption levels.
- According to this report [PDF], per capita water demand went from 158 gal/day in 1980 to 94 gal/day in 1990 to 111 in 1998 to 121 in 2003.
* LADWP's blocks appear to be this way -- they jump after 14 HCF/month and only by 15%. Even worse, houses on bigger lots get a bigger allowance at cheaper rates.