The inadequate supply of electricity that led to the recent rolling blackouts occurred because no one wanted to sell power to companies teetering on the verge of bankruptcy. Once producers have firm assurances that they will be paid, there should be enough power in California to meet current needs. The big squeeze will come in the summer, when peak demand can be up to 50 percent higher than in the winter.

Because it takes at least two years to build a power plant, the only viable solution for the coming summer is to manage demand more effectively. Offering rebates for purchases of energy-efficient appliances is certainly a step in the right direction. But realistically speaking, prices for electricity will have to rise a substantial amount to encourage conservation. California politicians are reluctant to push for a large rate increase because that could easily precipitate a host of citizens' referendums, adding to the confusion. But there is a way to break this logjam and set prices that will satisfy both economic and political objectives.

Economics teaches us that the structure of prices should reflect the structure of costs. Electricity generation is characterized by low marginal costs of production until plants are near capacity, at which point marginal production costs soar. One consequence of this is that it makes it both tempting and relatively easy for electricity producers to cut back output when they are near capacity, pushing prices up sharply. Some of this behavior seems to have occurred in California; but even if companies don't exercise market power, cost schedules of this shape can lead to very high prices.

The best way to deal with this sort of cost structure is with real-time pricing, in which prices are higher when the system is operating near capacity and lower during off-peak periods. This encourages users who are able to shift consumption away from peak periods to do so. The wholesale prices on the California Power Exchange exhibited exactly this pattern, but it didn't do a lot of good because the prices to end-users were constant. Real-time pricing requires new meters and new billing software, but many large commercial and industrial users could be outfitted by June 1.

In one plan now being discussed, industrial participation in real-time pricing would be voluntary. If a company used the same amount of electricity it used last year during the comparable billing period, it would pay last year's rate. If it used more, it would pay the much higher real-time rate on the additional use. If it used less, it could even get a rebate on the energy saved. The California economy was booming last summer, but it is likely to be more subdued this summer, making this plan potentially attractive to many commercial users.

It is not feasible to move residential users to real-time pricing in the next few months because of the time and expense of installing new meters, but there is an easy way to get some of the same benefits using a plan I suggested here in January. Most residential users of electricity in California now face a two-part tariff: they pay a low rate for use up to a baseline, then a higher rate above that.

It would make a lot of sense to increase below-baseline rates by a small amount and above-baseline rates by a significantly larger amount. In the electricity trade, this type of rate structure is known as an ''increasing block rate.'' Such a pricing schedule meets the economic requirement that consumers should face something close to the marginal cost of electricity, while still addressing the political constraint that average prices not rise too much. Consumers who choose to conserve would be rewarded with low bills; consumers who squander electricity would pay dearly.

The current definition of the baseline depends on climate, season, average energy use and other variables. This definition could be significantly improved. One choice worth considering is to set the baseline at, say, 90 percent of the amount consumed a year ago during the comparable billing period. Last year's consumption is printed on PG&E's bills now, so this should be possible to put in place in time for the summer. It is worth considering paying rebates to consumers who reduced their consumption below last year's level.

An increasing block rate would certainly help this summer, but it is just a stopgap measure. In the long run we have to have more generating capacity as well as more informative electricity prices.

No power plant applications were filed before the California Energy Commission from 1994 to 1998 because of the uncertainty surrounding deregulation. Unfortunately, while investors waited for the rules to be clarified, demand for electricity continued to grow.

A not-in-my-backyard mentality also played a role. Last summer, PG&E proposed putting a floating power plant in San Francisco Bay that would provide peak-period electricity for 95,000 homes, operating at most 200 hours a year. But the plan was scuttled in the face of opposition by local environmental groups. A proposal for a new power plant for Silicon Valley has been voted down twice, leaving the region dangerously short of power. Nobody wants a generator in his backyard -- until the lights start going out.

These delays are now a thing of the past: five new power plants will be online by the end of this year, with eight more starting operation in 2002. The biggest problem now is to get through the summer, and the only way to do that is with higher prices.