Alt-A loans were those made to borrowers who provided nothing more than their FICO scores (no income information, etc). It makes sense to believe that Alt-A's are hiding something and are more risky.
Subprime loans were to borrowers who provided sufficient information, and it was found wanting.

Presumably, the loan rates were set high enough to compensate for the risks. Homeownership would flourish, the real estate market would be robust.

When the going gets tough, the Alt-A's and the subprimes default at a much higher rate than do the standard loans. However, those who made those transactions do not get penalized-no claw-backs, although the institution suffers.

Moreover, the basic idea of mortgaged backed securities was that a mixture of mortgages provided greater certainty of returns than would any single mortgage, that is lower variance. The Central Limit Theorem, but it only applies if there is the random variables are independent, which in this case they surely are not.

The idea here is that MBS's allowed more investment funds to flow into the housing market, since investors now could treat mortgages much like any other security, with a risk and a rate of return. In general, this seems like a good idea. What was once a specialty market now is commodity.

But, those MBS's were inappropriately rated by the rating companies, with insufficient attention to problems of correlation and risk. Rates of return seemed to have little to do with the assigned quality of the MBS's.

ANALOGOUSLY: Alt-A dossiers have insufficient information, subprime ones are wanting. It may well be likely that these candidates will become dead wood with higher rates than would standard dossiers.

The idea that a departmental unit can absorb some of the Alt-A or Subprimes, using them in effective ways, works about as well as the MBS's worked. It makes sense to take on these risky individuals since they allow a department to grow in useful ways, yet the risk is controlled to some extent by the organizational effectiveness of chairs and deans. That is, rather than wait for stars to be hired, rare and hard to get, lets try a new technology, the organization or bureaucracy, that allows for quality growth without stars.

The ratings we quote about departments are perhaps even less reliable than are the ratings for MBS's

Moreover, deans and chairs get credit for hiring and promoting and for organizational growth (including research units, specializations in faculty focus (eg "string theory," "gender and dentisty"), but little claw back for mistakes--often they are out of position by the time the dead wood rots. It is a good idea to develop organizational inventions, but you need to know the risks involved--and plan around that. But, in general, it may be hard to know the risks fully, and in response to increase your tolerance for (now unknown) risk may be setting yourself up for a Meltdown.

There's no substitute for on-time delivery and performance, brains, and hard work. No organizational solution to a portfolio with lots of Alt-A and Subprime people.