Modern Risk Management- Time To Return To Yesterday?

Many years ago, mortgage finance was a regional business. Back in those days, pre-1980, the US commercial banking industry consisted of several large 'money center' banks- Chase Manhattan, Bank of America, Citibank, First Chicago, Chemical, Manufacturers Hanover, and Contintental Illinois, to name most of them- and many smaller regional banks. The latter would do business with the former using a model known as 'correspondent banking.'

In a sort of semi-formal partnership, smaller local and regional banks had an explicit relationship with their big-city, money center correspondent bank which allowed them to buy pieces of whole bank loans. Similar to the securities industry's syndication of underwritten listed securities, banks syndicated pieces of loans to their correspondents.

For the most part, due to restrictions on deposit-taking across state lines, money center banks sold loan participations to regional and local banks, thus, in effect, gaining by these sales funding which they could not access as direct consumer deposits in other states.

This correspondent function also had the effect, albeit, in this pre-electronic era, at considerable expense in terms of people and operating expenses, of diversifying loans throughout the US economy, on a loan-by-loan basis.

Particularly prevalent in these correspondent relationships would be business loans. The system was not without risk. Several banks, such as SeaFirst, in Seattle, went bankrupt after buying too much correspondent-originated energy loan volume in the infamous Penn Square Bank debacle of the early 1980s.

To see how regional mortgage finance was at the time, witness how many regional banks failed in the residential mortgage finance crisis of the late 1980s, when the RTC was created to clear the bad loans from the active banking system. Shortly thereafter, Salomon Brothers, First Boston and Kidder, Peabody dove into the private-label securitization of S&L mortgage assets, liquefying the balance sheets of the latter, and expanding capital available to housing finance.

Fast-forward to today.

In the last decade, the regional mortgage finance model was scrapped, as Countrywide Finance, Pulte, Toll Brothers, Centex, and other builders, helped to create a nationally-scoped residential construction and finance industry. With Congressionally-approved increases in the sizes of jumbo-mortgages taken for securitization by Fannie Mae and Freddie Mac, mortgage securitization reached new highs.

Investment banks, including Merrill Lynch and Bear Stearns, actually bought mortgage origination firms to provide a lower-cost supply of the instruments, with which to manufacture, for sale, securities backed by mortgages. Citigroup did likewise. Other firms, such as Morgan Stanley and Lehman, vastly expanded their mortgage securities trading and sales operations.

In effect, rather than restrict mortgage delinquencies and defaults to the regions which were typically economically impaired, and the banks in those regions, the new, securitization model of mortgage finance spread mixtures of mortgages from anywhere in the US into structured finance instruments sold to investors on a global basis.

Thus, when US real estate markets such as California, Las Vegas, and Florida, overheated, stalled, and collapsed, overextended mortgagees in those markets began to default. Instead of a few regional financial institutions being easily identified as affected, insolvent, and removed from the industry, the effects of the defaults were hidden amongst thousands of structured mortgage-backed instruments held in portfolios worldwide.

It doesn't take a genius to see how this insidious spreading of risk, meant to lower systemic risk, actually increased it by causing massive, ubiquitous concerns over counterparty risk.

To use risk management language, default risk of a particular instrument was replaced with counterparty risk involving any other institution suspected of holding mortgage-backed paper that might contain questionable mortgages which were about to default.

Thus, what began as a new idea for spreading and, thus, reducing, systemic risk from mortgage financing, while providing more capital, ostensibly at lower rates, to this sector, has, in fact, created disproportionately more risk, as counterparty risk, to the entire global financial system.

This brings me to a point I made in conversations with my business partner recently.

Way back in the early 1990s, when I was attending CMO conferences sponsored by Salomon, First Boston, et.al., I discussed the phenomenon with my boss, Chase Manhattan Bank SVP of Corporate Planning and Development, Gerry Weiss. How, I asked, could the total risk of a system which was attributed to mortgage finance, be reduced by adding more layers of securitization, with each layer taking a 'haircut' of a few basis points for expenses?

It simply makes no sense that total risk can be reduced. Any one player's risk may be reduced, but only by transferring, i.e., selling it to another party. The benefit to the system would be that each party could lay off excess risk, beyond its ability to afford or, ultimately, bear the consequences of the risk turning into actual loss. Thus, more parties could assume some risks, up to their capacity, in the form of mortgage-backed paper, to which they otherwise would not have had access, while other parties could reduce their risks to match their ability to sustain loss.

But, surely, the system did not reduce its risk.

Now, we see the outcome of this new approach. It doesn't work. Instead of containing losses, due to realized risks, among a few real estate-heavy lending institutions, these losses have spread among countless other institutions, be they underwriters, trading desks, or portfolio managers.

If, as we know from the last few decades of corporate finance teaching and research, an investor can more easily diversify his holdings for himself, than any corporation can for him, could the same be true for investors, with respect to fairly opaque structured financial instruments?

Would it not have been easier, when all is said and done, for institutional investors to pay a bit more to pick and choose the mortgages they wish to hold, rather than hope that someone who is paid for simply 'slicing and dicing' payment streams of mortgages like some financial butcher, without retaining risk in the structured financial paper he has manufactured, has accurately represented the risks therein?

Buying individual, seasoned mortgages directly from portfolios of banks or other financial institutions which underwrote the loans would have probably, in the end, cost less than what the current financial mess is costing us in terms of fear, uncertainty, and extensive counterparty risk.

I think it's time to acknowledge that securitization, as we have seen it for two decades, does not work. Unless the model is scrapped, or somehow modified to require significant shares of any securitization to be held by the structurer, who among us will again trust the party who sells us structured finance paper, being paid for the transaction, without holding it himself?

Maybe returning to the days of more regionally-based residential finance, geographical firewalling of mortgage finance problems consistent with economics of the region, and the buying of individual mortgages, or seasoned securitizations of them, makes more sense.