The Federal Housing Administration has suffered twin policy and financial failures. First, it is failing to help far too many first-time and financially constrained homebuyers become successful owners. Recent research projects that 15 to 30 percent of borrowers whose mortgages the FHA guaranteed since 2007 have defaulted or will soon do so. Second, it has failed to remain financially solvent. The FHA’s main mortgage insurance guarantee fund does not have sufficient funds to cover expected losses according to its most recent actuarial review, and a stress test showed that the FHA could lose $115 billion if the economy were to suffer a downturn.

The failures of the FHA call for a fundamental rethinking of how to better achieve the FHA’s mission: helping first-time homebuyers and financially constrained households with limited down payments become successful homeowners. The problem is that the FHA’s business model is fundamentally flawed. Both the FHA and the borrowers whose mortgages it insures are leveraged by more than 30-to-1, which is about the same level employed by Lehman Brothers and Bear Stearns before they failed. For this to be a viable business strategy for the FHA virtually requires that housing values never fall. We know from experience that markets are not so accommodating.

Replacing the FHA is far better than trying to reform such a fatally flawed operation. I recommend terminating the FHA by phasing it out over the next three to five years and replacing it with a subsidized savings program that would help targeted households build down payments for a future home purchase. The multiyear transition period helps ensure there is no sudden shock to the market, and significant policy reform like this is best implemented when the housing sector is recovering, as is now happening.

The new program would have several virtues relative to the FHA. Simplicity is one. My plan would allow only qualified households that the FHA is charged with supporting to pay into a special savings vehicle and receive a match from the government. The funds would accumulate on a tax-free basis until large enough for a 10 percent down payment on a modestly priced home, which would enable the household to obtain a mortgage without special FHA guarantees. This stands in stark contrast to the current system, which requires a large agency to price a complex mortgage insurance guarantee and manage a difficult foreclosure process.

Increased transparency is another benefit. If the goal is to help certain borrowers, the best way is to directly subsidize them so that they, not politicians or private market actors in the housing industry (such as homebuilders, realtors and lenders) are more likely to reap the program’s benefits. Having highly visible costs also allow policymakers to right-size the program by balancing costs and benefits. The efficient outcome is when the social benefit of the last subsidy dollar equals the social cost to the taxpayers who fund the program. Not being able to do this has been a major problem with Fannie Mae and Freddie Mac, not just the FHA, where the costs of opaque mortgage insurance guarantee and investment programs were “lowballed,” allowing huge risks to build over time.

Replacing the FHA with a subsidized savings program for targeted households also will generate far more sustainable homeownership. Having the government provide a match to household savings over time encourages financial discipline and long-term planning, in contrast to the highly leveraged gambles abetted by current policy. Unless one had the very good fortune to be born rich, sacrifice and deferred consumption always have been needed to sustainably afford an expensive durable good such as a home, regardless of the conclusions by some policymakers and bankers during the last boom that meaningful equity was no longer needed to buy a home. Because households will have substantial equity at purchase under the new program, they will be much less vulnerable to downturns in the housing markets, so substantially lower defaults can be expected.

Nothing is free, but this program would be less expensive than the true costs of operating the FHA over past cycles. No more than $1 billion per year would be needed to fund a very generous subsidy based on calculations I have made in other work. This is far below the $50 billion to $100 billion that recent analyses (including my own) indicate is needed to put the FHA on a sound financial footing. Perhaps even more importantly, a much safer and less leveraged housing finance system results in the bargain with a lower level of mortgage defaults and better personal outcomes for tens of thousands of borrowers. This is a win-win for taxpayers and financially fragile prospective homebuyers that the FHA is charged with supporting.

American housing policy failed utterly in the last cycle. It’s time for bold thinking about how to move forward to promote homeownership while protecting the homeowners and the taxpayers. Creating a responsible housing finance system by replacing the FHA would be a tremendous first step.

Joseph Gyourko is an adjunct scholar at the American Enterprise Institute and the director of the Zell/Lurie Real Estate Center at The Wharton School at the University of Pennsylvania.