Obama's mortgage plan under fire from left and right

President Barack Obama visits Erickson Construction on August 6, 2013 in Chandler, Arizona.
Mandel Ngan | AFP | Getty Images
President Barack Obama visits Erickson Construction on August 6, 2013 in Chandler, Arizona.

The most important question to ask about the Obama administration's proposals to wind down Fannie Mae and Freddie Mac and reform mortgage finance in the United States is quite simple: Will it work?

The basic outline of the plan is that Fannie and Freddie will get out of the business of insuring new mortgages. In their place will be some sort of private sector investment or insurance that will take the first loss on mortgage-backed securities, backed up by a government insurance agency that will cover any losses beyond what is covered by private insurance.

This is more or less what is called for in the bill introduced by Senators Mark Warner and Bob Corker. That proposal has come under widespread criticism, with both Peter Wallison of the American Enterprise Institute and Yves Smith lambasting the plan.

Wallison's main argument is that the political dynamics that led to the insolvency of Fannie and Freddie will simply return and recreate the same problems all over again.

A major feature of Corker-Warner is the requirement that the private sector share the insurance risk with the new FMIC. The bill specifies that a private risk-sharer like a bond insurer must take the first losses, no less than 10 percent on any securitized pool of mortgages. This is intended to protect the FMIC against losses, though it works only if the quality of the mortgages remains high.

But as with Fannie and Freddie, the quality of the mortgages will be the weak link. Realtors, home builders and community activists all want as many home buyers as possible. They are not concerned with fuddy-duddy obstacles like down payments, solid credit scores or low debt-to-income ratios. These interest groups and Congress will press the FMIC to lower its underwriting standards so that more and more loans can be insured.

The private bond insurer or other risk-sharer will understand the downside potential of low-quality loans and will charge for the additional risk. That cost will be incorporated into the mortgage, increasing the monthly payment and making the cost of mortgages too high for many potential borrowers.

The result? Congress or the administration or both will pressure the FMIC to lower its insurance fees so that the maximum number of people will be able to buy homes. Recall that the Federal Housing Administration required 20% down payments when it was born in 1934. It now requires a 3% down payment—and a government bailout

Yves Smith's critique is even more fundamental: she doubts there are investors willing to take the first loss position.

The other route might be to use bonds to transfer the first-loss risk. That sounds great, since risk would be transferred to well-diversified deep pockets. But in reality, this idea will founder for a different reason: Sophisticated investors are unlikely to want this paper, except on a very selective basis.

Remember, this is exactly how sub-prime worked. Sub-prime bonds were tranched by credit risk, with the riskiest tranches bearing the greatest risk of loss and therefore getting lower ratings (this was achieved by having they higher-rated tranches get priority in payment, and only when the money borrowers sent in every month met the amount due for the AAA tranches did any funds go to the AA tranche(s), and so down the ladder. The effect was that the lowest tranches bore the first losses. Now the very very bottom tranche (net income margin bonds, or NIM bonds) got extra goodies (extra interest margin in the deal, a loss buffer) so they were popular, since on a risk adjusted basis, if the deal worked out at all, they paid out fast and paid a premium return. So the real turkey was the lowest rated tranche, usually a BBB or BBB- tranche.

Here's the dirty secret: there was never enough of a market for that stuff. Both times the U.S. had a meaningful sub-prime market (a small one in the 1990s and the bigger one we know all too well), it depended critically on CDOs. The CDOs would buy the drecky stuff no investors wanted, plus some better stuff to make them more palatable. They were then tranched. But really no one wanted the bad parts of the CDO either (sometimes stuffees bought them, sometimes correlation traders would go long one tranche and short the one immediately above it) and so they were mainly rolled into OTHER CDOs. So the sub-prime market was ultimately a Ponzi scheme, and both times (late 1990s and the 2000s) both the CDO market and the sub-prime market imploded.

Let's say the Smith is too pessimistic here and there is actually a viable market for securities or insurance that would absorb the first ten percent of losses on mortgage-backed securities. After all, we appear to have entered a cycle of rising housing prices. There may well be investors who are—at least temporarily—bullish enough to take the riskiest parts of these deals at prices that don't make securitization impossible.

(Read more:Why is Obama on Zillow?)

But what will happen in the next period of financial distress? During the financial crisis, securitization was only kept alive by Fannie and Freddie. The private market was entirely shut down. In the next period of financial distress, why would anyone suppose the market for the first loss securities or insurance would remain robust? Most likely, it would completely shut down again. In that case, securitization itself would come to a halt since the availability of government backing would be dependent on private capital investment.

The probable result is that the requirement for private capital in the first loss position would just be dropped as "an emergency measure." The taxpayers would once again be in the position of backstopping nearly every new mortgage.

In short, we'd be right back where we are today.

—By CNBC's John Carney. Follow me on Twitter @Carney