Having just bashed the city's active water conservation program, let me put in a good word for mandatory watering restrictions.
I’ve never had anything good to say about command-and-control watering restrictions. Although they are much, much cheaper for the city than its active conservation program — and much more effective — they create dead-weight losses that don’t show up on the books.
There is a decent economic argument for watering schedule restrictions, though. I don’t remember anyone raising it in the comments to my many entries on this subject (although see Scooby's comment to the last entry), so I’ll lay out the case myself.
Prices are a great way to ration a scarce good. Usually. There are in fact times when command-and-control is better. Command-and-control makes sense when the marginal cost of supplying the good jumps sharply and dramatically at some point -- e.g., when it costs next to nothing to produce the first 100 widgets but the 101st is extremely expensive.
Consider a traffic intersection. We have to ration access to the intersection when several drivers want to enter at the same time. Why not use price? In theory, we could set up a transponder and debit every driver's TxTag account as he passed through the intersection.
This is utterly impractical, of course. And silly. Even if we could figure out a practical way to charge drivers for that intersection, allocating right of way by price is a bad, bad idea. Most of the time, perhaps, only one driver will want to enter the intersection. No problem. The price (= marginal cost) should be set a $0. But when two drivers want to enter the intersection at the same time, there is no reasonable price that will efficiently allocate right of way. There still might be two drivers who wish to pass through the intersection at the same time and who are willing to pay the price to do so. Then we have a wreck. Price alone won't do the job. It's much safer -- and cheaper -- to use stop signs or traffic lights. And these are just a type of command and control.*
Or think of a movie theater owner. The marginal cost to the owner of seating another moviegoer is $0. Until the theater is at capacity, that is. Once the theater is at capacity, the fire code prohibits the owner from admitting one more patron, even if the patron is willing to stand. If someone were to rat the owner out to the fire marshall, that last patron would cost him the whole kit and caboodle.
An omniscient owner would simply price his tickets so that there were exactly as many willing customers as seats. But in the real world, owners don't know how to set that price. Part of the problem is that attendance is random. Sure, the owner can predict more moviegoers on Friday than Wednesday, and more for Transformers II than Away We Go. But traffic fluctuates, so the owner can't be sure that everyone won't wake up one Saturday and say, "Let's go to the movie tonight." If the owner charges a high enough price to deter random spikes in attendance, he'll be overcharging patrons on "normal" nights. So the owner charges a uniform price and simply turns people away when the theater sells out. Again, this is a form of command and control.
In both examples, the cost of supplying the "good" is very low until demand reaches a critical threshold, when the cost shoots up suddenly and dramatically. And, in both cases, the demand for the "good" has a significant random component.
Peak-day water demand is arguably like this. Suppose that Austin Water Utility's absolute maximum daily capacity is 275 million gallons of water. It can supply the first 275 million gallons of water at a marginal cost of a fraction of a cent. But once it's supplied 275 million gallons, it doesn't have any more. If demand reaches 275 million + 1, then AWU must either cut everyone off or it must truck in water. Neither is an attractive option. AWU's supply curve thus has a kink at 275 million gallons.
Demand for water is also like demand for traffic intersections and movie theaters in that it has a random component. An intersection can accommodate all vehicles as long as they don't try to get through all at once. A movie theater can seat all moviegoers as long as they don't try to go the show at the same time. And AWU can supply enough water to meet Austinites' daily needs as long as we don't all decide to soak our lawns on the same day. But just as an intersection will see random spikes in traffic, and a theater will see random spikes in attendance, there will be days when Austin's water users go hog wild with the water at the same time.
Pricing alone is not the optimal solution for water just like pricing alone is not the optimal solution for the theater. AWU could charge a high enough price to deter a devastating spike in peak demand. But this means setting the price high enough to keep water use low even if everyone by chance decides to water their yards at the same time. That would be a very, very high price. Too high a price, because residents won’t usually all water on the same day. On “normal” days, they will have to pay an exorbitant penalty.
So water users are better off with watering restrictions because the restrictions smooth demand and allow a reasonable pricing scheme to get the job done.
Note that this argument does not justify mandatory watering restrictions as a means of conserving water over the course of the year. Even in a scorching summer, prices are more than adequate to hold annual consumption down to a reasonable level. Prices work well when we’re rationing stock; the above argument shows they work less well when when we’re rationing flow.
*If you want to get cute, we arguably do ration access to intersections via price. A driver who runs a red light pays a price equal to the cost of a ticket times the probability of being caught plus the expected cost of an accident. Drivers systematically underestimate the latter “price,” which is why we have so many deadly accidents at intersections.