New York Gov. Andrew Cuomo has requested $30 billion worth of federal aid to help rebuild after Hurricane Sandy. New York City needs reconstruction, but our current federal relief system leads people to build excessively in high-risk areas and provides little incentive for battered states to make wise decisions about which reconstruction efforts are worth the cost.

To promote a more sensible recovery, Congress should link the aid New York needs to a change in the way the Federal Emergency Management Agency operates. Specifically, FEMA should split its emergency-response functions and its insurance functions into separate entities.

FEMA currently does two jobs. First, it protects life and property in the heat of the moment, as a fire department would. Second, after a disaster, it steps in with two large insurance programs — the National Flood Insurance Program and the Disaster Relief Fund — which together spend about $10 billion out of FEMA’s $14 billion budget.

Since disasters don’t strike every part of the United States simultaneously, FEMA’s emergency response capacity helps coordinate out-of-state response resources and move them into the affected area. A national agency has the system-wide perspective to promote investment in emergency-management infrastructure and assess far-flung risks. Hurricane Katrina exposed the weakness of one state’s local response efforts; FEMA can ideally ensure that every American is defended against natural disasters.

There is also a strong case for a national insurance program that helps share the risks associated with rare expensive natural disasters. Private insurance is fine for small, idiosyncratic risks, like car crashes, but it’s hard to imagine private insurers covering the vast costs that a natural disaster can inflict on state governments. Thirty billion dollars for New York State would require almost one-fourth of Allstate’s total assets. Insuring state governments is also inherently problematic, because those governments enjoy regulatory powers that would enable them to practically dictate payment terms.

But while national insurance makes sense, it currently causes mischief. The most famous insurance problem is moral hazard, which refers to the tendency of insured individuals to take inappropriate risks. If my car is insured against theft, I may park it on the street of an unsafe neighborhood. When the federal government insures homes and states against hurricanes, there will be too much building in hurricane-prone areas.

FEMA’s Disaster Relief Fund creates a second, less obvious, problem. When your Honda is totaled, the insurance company will typically give you a flat sum, so you can buy a new Civic or anything else you like. When a hurricane destroys infrastructure, FEMA provides up to 75 percent of rebuilding costs, whether or not rebuilding makes sense.

Cuomo’s request for $30 billion probably isn’t unreasonable; after Hurricane Katrina, FEMA and private insurers spent far more to clean up and rebuild areas of the Gulf Coast with far fewer residents than New York. But as long as FEMA is paying the bill, Cuomo will have little incentive to forgo wasteful projects or contain costs on necessary ones.

Ideas from the private insurance market can mitigate the current problems with disaster relief. The moral-hazard problem would decline if we adopted risk-based insurance premiums, where different states would contribute different amounts based on the risks they faced. If a state directs construction away from a hurricane-prone coast, its premium would fall. We should reduce the subsidies that FEMA’s flood insurance program provides for wealthy homeowners living in risky coastal areas, and also tie the premium more tightly to the expected insurance costs in specific areas.

When rebuilding, states should have the option of project-based rebuilding assistance or unrestricted cash, so that, like car owners, states and localities can choose whether and where to rebuild.

FEMA’s emergency response team served us well during Hurricane Sandy, but its insurance arm needs an overhaul. An independent insurance entity, funded by risk-based assessments on states and individual homeowners, will cause less moral hazard. If the insurer pays with unrestricted cash, communities will do a better job of rebuilding only when the benefits exceed the costs.

Edward L. Glaeser, a Harvard University economist, is the director of the Rappaport Institute for Greater Boston.