Editor's note: This is Part 2 of a special series. Here are Part 1 and Part 3

NEW YORK ( TheStreet) -- "What is a triple-A?" The rating agencies look at their long-term historical default profiles of triple-A investments. They may differ slightly in whether or not they are rating to the first dollar of loss or the expected loss profile of a triple-A, but this is essentially what they are seeking. So they look at ratings migrations, jumps to default, etc. of their histories rating assets. They may determine, "OK -- less than 0.01% of triple-A-rated assets default over a 10-year horizon." So if they are going to rate a structured finance instrument, they want to make sure that it has less than a 0.01% chance of defaulting over a 10-year period.

So how do they do that? This is where they analyze the historical performance of similar looking mortgage pools. Here they might say, "The typical MBS pool experiences a 2.5% loss rate, so we want you to cover six times that amount to get a triple-A, and we want to run several scenarios front-loading and back-loading these losses to see that the tranche you want rated survives to triple-A standards." Furthermore, they release their ratings methodology so people can see to how they got there.

And the institutions that play in this market have their own portfolio managers and analysts who do their own modeling with their own assumptions and their own stress tests. If you play in this market, there are loads of analytics providers that cater to the evaluation of these securities.

> > Bull or Bear? Vote in Our Poll

But there is a little bit of a "framing" issue here. If the historical loss averages 2.5% for these pools, then one is not likely to think 20% is even in the realm of possibility. And that is a problem, particularly when there is a structural shift in how mortgages get originated.

I am not sure if it was causal or coincidental, but starting in the 1990's when the nonagency mortgage market began to pick up pace, so too did the way mortgages were originated. Maybe it was because banks no longer kept their loans in their portfolios and used securitization to get them off the books and refresh, but I also suspect it was the development of "the food chain" of the mortgage process.

We saw a shift in the way the process worked. Instead of just going to a bank or savings and loan (many of which were just coming out of the last housing crisis, so probably had tighter lending standards), people began to go to mortgage brokers, which helped borrowers "shop" for the best loan. The mortgage broker filled a need. And the regulators and government officials thought this was great because it was the "free market" helping out the consumer.

The mortgage brokers would then in turn, sell the loans to a mortgage aggregator. Maybe it was a bank, or maybe it was a nonbank financial, that would buy from that broker and many others and assemble a pool of mortgages. Once they had critical mass, they would sell that portfolio to a Wall Street firm, which would securitize the pool into bonds and sell them to institutional investors.

This was the free market in action. The consumer benefitted by getting a competitive rate, which was helped by government policies that incentivized the consumer to take on debt with a tax break, and every link of the food chain was able to take a nibble. (This is the part that many missed or were blind to, but Paulson and Scion and a small handful of others recognized what was happening.)

But what the government officials ignored (and this is Lesson of the Crisis No. 1) was that they were ceding control of the credit creation and circulation process in this country. Thus, even if they wanted to, they didn't have authority over these origination processes. So Lesson No. 1 should be: If you want to become part of the credit creation process and recirculate credit, guess what, you should be subject to regulation. So many of these folks were not, and that helped things spiral out of control.

What's worse, the government and regulators were somewhat complicit. The increased competition from outside the banking sector caused the banks to degrade their own lending standards. After all, it was a regulated financial institution after all that created the "pick-a-pay." But these banks told their regulators, "We will not be able to compete if we don't do this." And reluctantly or not, the regulators said, "OK."

By ceding control of the credit creation process, the regulators made a huge blunder. Particularly when we further used tax incentives to get borrowers to drink from the debt trough.

The lending products became more and more "interesting": pick-a-pays, no-docs, stated income, etc. The whole time we were being exhorted by Congress to let the little guy be a part of the housing action. (And as in Casablanca, we now have authorities who are "shocked, shocked" to find out what happened.)

This digression on the shift in origination actually has big implications for the evaluation of the rating process. No longer are the underwriting standards at the point of origination the same. So looking at past history is irrelevant. At first it was gradual, but then as we got to the blow-off top, we realized something was amiss with performance vs. expectations.

So if we want our triple-As to be able to withstand six times a normal loss rate, what happens to those if we change our opinion of what "normal" loss rates should be? If we move from viewing 2.5% to be normal to maybe 5% being normal, we have a leverage multiplier here. Now we only have three-times coverage, and our triple-A no longer looks so solid. Maybe it is a triple-B instead. There's still plenty of coverage, and it's still an investment grade rating, but think of all of the securities out there and think of the proportion of investors who are only allowed to hold triple-A assets. One change in a modeling assumption has a dramatic effect.

This is an issue with no great solution. But when I hear that 70% of the bonds in a mortgage CDO pool are downgraded within seven months of issue, I think that it was due to a regime shift at the rating agencies, such as going from 2.5% to 5% as a "normal" loss rate. Should they just rerate on a prospective basis and leave extant deals alone? Should they not change opinion? These are hard questions without easy answers. This is further complicated by the fact that more than half of rating agencies' revenues were driven by structured finance instruments.

These are also hard questions because of the amount of influence the rating agencies have. At one point the Basel bank accords wanted to consider ratings based capital standards. What happens then with a regime-based ratings shift? All of the sudden a bank is dramatically undercapitalized.

And although nearly every institutional investor will tell you they do not rely on the rating agencies and that they perform their own analysis, they all have investment guidelines that say they will be in triple-A assets, or investment grade assets, or that below-investment grade will only comprise a certain percentage of their portfolio. Think about it. Do you have money in a bond fund? Is it an investment grade bond fund? Is it a high-yield bond fund? If you are on a pension or endowment committee, do you say, "Buy whatever you want"? Or do you say, "Let's stick to investment grade assets"?

Eric Oberg worked in fixed income, currencies and commodities for Goldman Sachs for 17 years before retiring as a managing director.

More from Opinion

Elon Musk's Latest Twitter Tirade Is the Dumbest Thing on Wall Street

Elon Musk's Latest Twitter Tirade Is the Dumbest Thing on Wall Street

Elon Musk's Twitter Tirade Is the Dumbest Thing on Wall Street

Elon Musk's Twitter Tirade Is the Dumbest Thing on Wall Street

Why Google's Search Momentum Won't Be Badly Hurt by New EU Rules

Why Google's Search Momentum Won't Be Badly Hurt by New EU Rules

Flashback Friday: Amazon, Chip Stocks, Memorial Day

Flashback Friday: Amazon, Chip Stocks, Memorial Day

Time to Talk Tesla: What Happened This Week, Elon?

Time to Talk Tesla: What Happened This Week, Elon?