Orphaned Mortgage Article

I had a mortgage article I was hoping to finish off with someone more credible and publish properly, but it wasn't to be.

Here's the draft.


Slicing the Wrong Way: Securitization could be done right

Securitization is not a bad thing. It is a good useful thing. We do not have a financial crisis today because of securitization. We have a crisis due to poor oversight, greed, stupidity, and perverse incentives. And we will continue to have them. Human nature loves to build castles in the sky. Bubbles are here to stay, all that can be done is to make them obvious enough for the prudent to avoid, and that good investments are not damaged by the bursts. We once had railway manias, now that we understand the rail industry, it no longer excite us. Mortgage markets need to be tamed and boring, rather than disappear.

Structured finance has been used to create illusions of value. But the illusions stem from the modelling and rating of those securities. Opaqueness hid risk, not the split between underwriter and investor. Institution's exposure to their own poor underwriting, suggests that encouraging banks to lend on balance sheet would not avoid our current problems. Here in the UK, Northern Rock could well have survived if more of its portfolio had been sold on.

The Great Depression was triggered by the total seizure of the global banking system. The first Black Monday in 1929 was private tragedy for stock market investors, which in itself did little to damage America's economy. Traded securities can only go to zero, and the investor has already parted with their money. Banks are different in being able to write credit as money. Under strict controls this provides necessary currency to the economy, and as banks would fully cover their commercial risks with their own capital. In reality, these controls are often not seen until long after they are needed. The over-extension of banks should be the main worry for the global economy.

Bank loans tend to rot in private, only becoming issues when the lender's solvency is already impaired. In contrast, stock bubbles happen in the open, where the buyer knows, or should know, the risks. Orderly price corrections require both transparency of asset values, and the ability to shift ownership of impaired assets. Open markets provide both. When failed dotcoms collapsed in price, they were swiftly asset stripped by vultures. Resources were freed up, and the economy carried on.

Mortgage are simple arrangements, accurately understood by hundreds of millions. Mortgage backed CDO are complex financial instrument, accurately understood by about eighty thousand people. Mortgages are pooled into a holding company, and that company issues differing types and seniority's of debt, called CDOs. The process of separating out credit risks into different form of securities is known as tranching. Tranching makes loan pools "fit" the risk models of potential buyers, especially conservative institutions. A normal CDO loan pool is a machine to turn mortgages [simple to understand] into AAA-rated bonds [too safe to bother understanding].

But a pool of one assert, cannot be easily treated as a single unit of another. For example, some mortgage backed pools will try to concentrate the risk of homeowner prepayments into "companion bonds". Of course, if the entire pools was prepaid, there is no way to offset this. Non-companion bonds are still exposed to mass prepayments. Or an insurance company might be paid to guarantee the safest bonds. Which is not so useful if the insurer's solvency fails with the bond. (As you would expect, and did, happen.) Extreme market conditions radically alter CDO behaviour, far more so than the underlying mortgages themselves. Even in good times, CDOs often fail to behave like conventional bonds. A small number of unexpected defaults can impair the most senior tranches by a tiny bit. How often do companies or governments become a 'tiny bit' insolvent?

Pooling then tranching mortgages, also obscures their title. In an extreme case, lack of proper documentation prevented Deutsche Bank from foreclosing on 14 Ohio homes. An openly traded security does suffer from fragmentation of control, but modern settlement systems can ensure accurate titling, while the ease with which ownership can be transferred compensates for a lack of single owner. Holders of a failed company can sell their rights to professional vultures in an ordered way. Vultures step forward to make offers whenever traded prices are at a discount to what they believe they can recover. They offer the market a single responsible investor "on-demand". Lacking clear ownership, the mortgage is effectively under the servicing company's control, rather than the investor's. Servicing companies have little incentive to rehabilitate mortgages, and often have perverse incentives (right of first refusal, for example) to foreclose.

Despite the complications, tranching an asset is often economically efficient. Just as a bank manager will want to see "hurt money" ahead of making a business loan, conservative investors will want to see someone else offering to take the first loss on a mortgage. This is rational, but must surely make more sense applied to individual mortgages, rather than diversified pools.

Suppose the existence of an open exchange for individual mortgages. Each mortgage is listed as a single entity. Each tranche of that mortgage is an individual security against a single entity. So a typical mortgage might become a thousand senior and a hundred first-loss $100 bonds. Could this work? I think it could. The valuation models already exist. Computers have made order processing costs almost zero. There is no logistical problem in having market with millions of different securities, it is just more 1s and 0s. (Stranger markets have been built. Sears operates an auction market for goods trucking. There are markets for surplus food, and shipping capacity.)

Liquidity on such an exchange would behave differently to a conventional stock market, but it could easily be just as high. There is no reason a trader should not make individually priced bids on ten thousand mortgages, capping his purchase at the first $2m of accepted bids. Such "broad" offers would allow the market to ebb and flow with macro economic changes, despite consisting of claims over very small idiosyncratic assets.

Although the seeing your mortgage's value shift every few hours might be disconcerting, the reality is that it does this anyway, so you might as well be aware of it. Price transparency protects both buyers and sellers. Mortgager can see accurate rates for their specific type of mortgage and property. It sets a market price for prepaying an "non-prepayable" mortgage, e.g. the cost of buying and voiding part of a mortgage. Existing owners could refinance cheaply and easily, knowing that the terms they receive will not be tied to their original mortgager. Defaults would be recognised immediately, and the vultures could move in to rehabilitate (or foreclose) such loans.

There are a number of objections that could be made to such a market. The most important ones would be: fear of market manipulation hitting individual owners, and financial privacy. Unlike shares, ownership of a mortgage does not generally give the investor any control over the payer. So were a household's mortgage bought out, this would not affect them if they were performing on the loan. A genuine potential problem would be vindictive short-selling aimed at undermining a household's access to credit. Also of concern is the handling of household credit data, which would be necessarily shared among thousands of people. (Although, in many jurisdictions mortgages are publicly registered. And an incredible number of firms have access to consumer credit records!) Any scheme for establishing a mortgage market will require strong internal oversight to let homeowners feel comfortable.

There are only three indirect precedents for markets in individual credits or properties. A market in the equity of individual residential properties, called Opromark, operated successfully in Britain for a few years. Opromark allowed purchases of as little as £1 of building. However it failed to find needed venture capital after the UK's buy-to-let market slowed. Another site, Propex, which deals in commercial property is still going strong.

The most significant is Prosper.com in the US. It allows private individuals to offer credit directly to other pseudonymous individuals over the internet. Borrowers make eBay-style listings requesting a loan. Lenders then bid in an auction to supply funding. (I helped build something similar in the UK [Zopa affliate link goes here!], but it does not yet have Prosper's operating history.) Prosper publishes extensive transaction level market data, so its behaviour can be carefully tracked. Initially, Prosper lenders suffered their own subprime disaster, funding many high risk loans that soured (or were fraudulent, the story of Jessica Wolcott is especially entertaining). But the Prosper market wrote that bad debt off, massively adjusted its lending criteria and continued to invest. The bad loans altered behaviour, but they did cause a general contraction of Prosper lending to good credits. The bubble inflated, popped, and normality quickly returned.

Currently, the Western public demands a rapid return to cheap credit, paid for by someone (preferably foreign) else. I cannot imagine the introduction of further market mechanisms being popular. Most likely they would be seen as an attempt to throw mortgage rates into a vicious "stockmarket casino", e.g. rise. (In the UK, this is going to cause crushing economic problems as banks are slowly forced to digest their bad loans in pseudo-Japanese fashion.) Besides significant capital could only flow into individual mortgages if the investment-grade straight-jacket on institutional investors were loosened. Something that can only realistically happen after everyone has had a chance to calm down.

What could easily emerge, is an institutional market for flipping single whole mortgages. As the PhDs and MBAs pick up the pieces, there will be a constant need to quickly sell up mortgages that do not fit with a bank's new strategy, and to buy in mortgages that do. As a insider's market it would not need new regulation, or be subject to public disapproval. And in time it could grow into something bigger.

Developing markets in Asia would probably be more comfortable with markets in individual mortgages (or other credits). Banks there are weaker, younger institutions than in the West. (Although over the last decade, Asia has improved banking supervision, and the West has degenerated.) And most of all, change is just easier to sell in younger markets. I have met 20-something Chinese saying "I remember when I was young and this was all trees", gesturing at a busy central shopping streets. For people only recently used to formal banking, any reasonable change is not going to seem too strange.

Proposals to return assets to lender balance sheets will only create more uncertainty. It is not even clear that banks have that capacity without further government bailouts. Covered bond program might enhance safety by adding a bank's guarantee to the assets. But an extra wrapper of credit increases uncertainty in valuations, even if it reduces the risk of loss. And the credit of the banks keenest to sell their loans, might well not be a major improvement over the asset's own. Selling on bad debts to clean up bank books is an accepted practice from dealing with China's bad debts in the 90s, to the S&Ls bailout in the US.

We are all globally reliant on market-based financing, and should not allow shady practices in US subprime to slow progress in an area with so much promise. Attempts to reform credit markets should look forwards to the more efficient and transparent credit markets that new technologies can enable.