The recent power outages in Washington have steamed a lot of people, literally and figuratively. An outage like this one reminds us how much our lives have come to depend on a reliable supply of electricity.

While the blackout has been annoying, imagine what it would be like if we lost power for weeks across several states. Impossible? Not at all. In 1998, that happened over a wide swath of eastern Canada when an ice storm knocked out power to 1,673,000 customers. In many cases the power was off for weeks, causing economic losses estimated to have exceeded $1.5 billion. A hurricane, ice storm or terrorist attack could do the same thing to us. If we were unlucky, power could remain off for a month or more across a region of many thousands of square miles.

An outage of that magnitude would be more than inconvenient. It would disrupt a wide variety of critical services, such as gas stations, supermarkets, emergency responders, schools and, in many cases, even basics such as water and sewer service.

Because the power system is spread out across the countryside, there is no way to make it 100 percent reliable. But three technologies could be combined so that critical services wouldn’t be so vulnerable.

The first is “distribution automation.” If a tree comes down and takes out a power line, this technology enables the power company to reroute the electricity to customers whose lines are still intact.

The second is smart meters. If the power supply is limited, smart meters could be used to connect only those customers who are providing crucial services. They could also be used to cycle limited amounts of power among a number of customers, so that at different times, everyone could get at least a little bit.

But in a big blackout, where would that little bit of power come from? That’s the third technology: Suppose places such as shopping centers, housing developments and industrial parks had their own medium-size, privately owned power plants. Ordinarily, these would supply their owners with heat, cooling and some electricity. During an outage, some of this power could be used to keep critical services operating elsewhere in the community.

Not all critical loads would be best served in this way. For example, it would be better to keep traffic lights going with batteries charged from solar cells (a technology that is already in use in some places). Major facilities such as hospitals would continue to depend on backup generators.

We recently showed in a paper in this month’s issue of the journal Risk Analysis how all of this might be accomplished. But three things are keeping it from happening.

First, to be cost-effective, privately owned “distributed generation” must be allowed to supply several customers through a “microgrid.” But most states (including Virginia and Maryland) and the District have laws that make it difficult or impossible for anybody but a power company to sell electricity. Thus it is illegal for such an operator to sell even small amounts of power to anybody other than the power company. As a result, these systems, which exist in other countries, aren’t being built here.

Second, while power companies are putting in smart meters and installing distribution automation, they have no incentive to make the small additional investments in control, switching and safety equipment that would be needed to allow a distribution system to provide limited power to critical services.

Finally, even if a utility wanted to install such a system, at present it has no way to pay for it.

If we want to ensure a reliable supply of electric power to critical services, we need to change the law so that private plants can easily get approval to operate small local microgrids. We need the U.S. Energy Department and others to fund a few pilot projects to show how systems could be built and operated to supply critical services. And, finally, to address the broader funding issue, we either need public utility commissions to declare the modest additional cost of such systems to be a “prudent investment” so utilities can afford the changes, or we need to start paying for those changes through tax revenue.

Our critical services don’t have to be vulnerable to large blackouts that last a long time. The technology exists today to eliminate this vulnerability. Policymakers need to make it happen.

M. Granger Morgan is head of the Department of Engineering and Public Policy at Carnegie Mellon University. Anu Narayanan is completing a PhD in Engineering and Public Policy on making power systems more reliable.