I’m not saying you’re wrong, but the single biggest consumer of water in the American West is agriculture, by a long shot. The ag sector is also probably the place where you can get the biggest bang for you buck out of conservation efforts - instead of convincing people one household at a time, you can go to land managers or irrigation districts and implement practices that affect huge areas of land and volumes of water. Historically, the USDA Natural Resource Conservation Service has done this by working with producers to develop and fund water conservation, and these projects have been continued in the last two farm bills (2002 and 2008).
On the other hand, those conservation practices don’t always have the intended result. If you save a producer 5 acre-feet of water that were previously lost to infiltration, that doesn’t return 5 acre-feet to the river; it means he can grow alfalfa instead of a crop with lower water demand. It also means that that water, which used to percolate back into the water table, doesn’t get there anymore. The end result of the whole project is a happy farmer and a drier environment.
Ultimately, the main problem is what Jophiel said: there’s just not enough water in the West to support the current populations, let alone growth. The Colorado River, for example, is fully allocated - on an annual basis, 7.5 million acre-feet go to the upper basin, 7.5 million acre-feet go to the lower basin, and 1.5 million acre-feet have to be delivered to Mexico per a treaty. Unfortunately, that allocation was based on a particularly wet year (1928, maybe?) following several consecutive wet years. Today, the Colorado River flows average between 13-14 million acre-feet/year.
That doesn’t stop development or planning based on the existing allocations, though. The shortfall can be temporarily buffered with reservoirs, but Lake Mead and Lake Powell (the two biggest reservoirs on the Colorado) are both half empty, and there’s no water on the horizon to fill them. Groundwater pumping on the scale that would be required to make up the shortfall won’t swing it, either. Groundwater and surface water aren’t really separate - if you depress the water table enough, you’ll disconnect the rivers from their base flows, and you end up with even less surface water. There are other significant potential impacts to groundwater extraction as well, including land subsidence, drying of existing wells, and impacts on endangered species habitat.
We can’t just create water. Desalinization has potential, but Chronos is right about the energy demand. The Department of Energy is funding some work in coupled energy-desalination processes, and some coastal areas have pilot scale stuff in the works. Aside from that, the reduction of consumption is really the way you have to go, either through reuse and recycling or just reduction. But the temptation is there to say, “well, we save a million gallons a day here, so we’ll use that water on this other project.” We can’t do that, because “that water” doesn’t physically exist. It’s a legal construct, it’s considered to be somebody’s property (specifically, it’s the property of the state and allocated to the holder of the water right), and it has a monetary value, but it ain’t wet.
Without attaching a cost to the use of water (rather than a one-time cost for the acquisition of water rights), there’s no realistic incentive for that kind of use reduction.
On the stimulus side of things, I don’t water infrastructure as a big job-maker. The state of the technology require research, not huge public works projects. With the way we’re managing our water now, we’re not going to create any jobs for construction or operations - we’re going to be creating work for lawyers.