This is a nice outline of the incentives for securitizing debt instead of making traditional loans:
The Treasury Department's plans for securitization reform are being bandied about in the press. A key question is whether it can or will fix the now-broken private securitization process.
Credit became more dependent on securitization than many realize. By pretty much any metric, the role of banks relative to other players has declined since 1980, by some measures as much as a 50% drop in market share. Securitization, which is the process of putting loans into pools and often slicing and dicing the cash flows to create instruments that are more appealing to investors, has been the big culprit.
In case you missed it, securitization has slowed down to a trickle. In the US, non-agency securitization was $900 billion in 2007, $150 billion in 2007, and a mere $16 billion through April. Now some of that was dodgy CDOs and other subprime spawn that is better off not coming back. However, if the securitization machine remains impaired, the alternative is on-balance sheet bank lending, and the authorities do not appear interested to going back to banking circa 1980. As the Financial Times notes today:
Securitised markets – which financed more than half of all credit in the US in the years immediately preceeding the crisis – are essential for the US economy. Without a recovery in these markets, the flow of credit will not return to more normal levels, even if US banks overcome their problems.
The reason securitization became so widespread is that it is cheaper than on-balance-sheet bank lending. Traditional lending requires banks to recoup their cost of equity and FDIC insurance premiums. For assets that can be packaged, securitization is more attractive.
From a policy standpoint, therefore, the desire to restart securitization is two-fold. First, it in theory produces more abundant and cheaper credit (although any reduction in yield depends on whether the banks and other participants keep all the cost savings in the form of increased profit or pass some of them on to borrowers). Second, going back to old-fashioned lending wold require banks to have much larger balance sheets, hence more equity. The banks are having enough trouble coming up with enough capital to support their current footings that raising even more equity would seem to be a non-starter.
However, it is not clear that the ideas floated by the Treasury will do the trick. It has two components: the first is requiring that the party that originates the loans to be securitized retain 5% of the deal. The second is to eliminate gain-on-sale accounting, which increased the attractiveness of securitization considerably. Again, from the FT:
The authorities plan to force lenders to retain part of the credit risk of the loans that are bundled into securities and to end the gain-on-sale accounting rules that helped spur the boom of the markets at the heart of the financial crisis...
The Treasury plans to force lenders to retain at least 5 per cent of the credit risk of loans that are securitised, ensuring that they have what investors call “skin in the game”. The 5 per cent rule – which looks set to be applied in Europe as well – is less draconian than some bankers feared. The proposed elimination of “gain on sale accounting” is to prevent financial companies from booking paper profits on loans – packaged into securities – as soon as they were sold to investors.
Banks would only be able to record income from securitisation over time as payments are received. Brokers’ fees and commissions would also be disbursed over time rather than up front, and would be reduced if an asset performed badly due to bad underwriting.
The US authorities also plan to stop credit rating agencies from assigning the same types of ratings to structured credit products that are assigned to corporate and sovereign
The proposal to change accounting and allow for clawing back of profits if a deal goes bad is probably far more meaningful than having banks retain 5%. 5% is simply not significant enough, in and of itself, to change behavior much.
However, there is an unrecognized contradiction here. The reason securitization became pervasive was both a real improvement in economics (bona fide cost savings per above) which were then compounded by efforts of banks to streamline costs further by scrimping on vetting loans (why bother if all you needed to on-sell the stuff was FICO scores and other simple metrics?). And the favorable accounting also was a considerable impetus.
Thus any activity that changes the incentives meaningfully will also reduce the attractiveness of the economics to banks. Some of this is salutary and necessary. The point is to discourage banks from selling dreck. But effective measures may reduce the size of the securitization market more than the powers that be anticipate. There may be no free, or even cheap lunches here. For instance, it might take a more meaningful retention (20%? 30%) to change originator incentives, but a proportion that lare would make securitiation a far more marginal activity and might require the move the powers that be are hoping to avoid, namely considerably more on-balance sheet lending.
Personally, I'd stick with the changes in accounting treatment proposed, but would increase liability considerably in the event of deficient due diligence and mis-selling (ie, burned investors could go after banks and rating agencies tooth and nail). However, it would take some effort and thought to come up with the right framework.
Sadly, the US seemed able to do that in the Great Depression. The provision of the securities laws of 1933 and 1934 were astute and durable. I wonder why devising good regulatory regimes has become a lost art.