In defense of Securitization

Credit crisis is about to claim its latest victim: securitization. This mechanism is now blamed by some as one of the important causes of this crisis. Lately, America’s most influential economist, Paul Krugman, was very harsh on securitization and the government’s efforts to revive it. He noted that “Bear Stearns failed; Lehman failed; but most of all, securitization failed”. I believe Krugman is making a big mistake here.

There is no doubt that securitization has been an enabler for banks to take ever larger risks. Yet, securitization is just a tool, like many others. It brings benefits, but it has its costs and risks as well. Using securitization to add more leverage, the bank managements were missing the larger picture. Systemic risks posed by this mechanism were not understood, even when the individual portfolios of securitization products were risk managed in excruciating detail. There was very little questioning of the systemic risks! What would occur in a dynamic, interactive system when everyone follows the same strategy. What was the extent of liquidity? How could you take it granted that markets would trust the accounting of banking books ignoring market realities?

Such risks of systemic nature are not unique to securitization. Every financial product, every financial mechanism is subject to abuse and misuse. Even good regulatory systems are no guarantee of stability. Paths of financial crashes follow the same contours: ever increasing risk exposures in the face of diminishing margins, overcrowding and eventual collapse. Consider Iceland’s financial crash in late 2008, for example. That has been achieved thanks to no help from securitization markets. From Asian crisis, to 2000 tech crash, we observed the same: the rush of capital to an increasingly crowded, competitive market, followed by the inevitable flush. Just as it would have been wrong to stop globalization, to stop influx of money to internet ventures at those times, it is wrong at this time to stop the securitization engine.

Let’s remember that securitization was only one part of the credit bubble. Our credit risk taking was not limited subprime mortgages. It involved prime mortgages, commercial mortgages, company debts, levered buyouts and acquisitions that inflated the equity markets and a lot more. Even without securitization, we had reached to an over-levered financial system primed to burst in short order.

Let’s consider also what would happen if there were no securitization. A large part of the world lives under such constraints. Obtaining credit in such countries is difficult or impossible. Financing is largely done by individual savings. Getting a house, financing a new business venture is prohibitive. Existence of securitization in richer countries has benefited their borrowers.

Just to put a number on it, I have made some calculations. I have found some figures on the debt accumulation of households and non-financial corporations in United States. The data I found indicates $16 trillion cumulative borrowing by them since 1995 (approximately the point when securitization has accelerated). I assume 50% of this borrowing has benefited from securitization. Here is my critical assumption: I suppose these borrowers may have paid 2%-3% less on their loans they would have otherwise, thanks to securitization. Assuming 2.5% less interest rate on the loans for every year, I guesstimate the total reduction in interest paid by these borrowers might have added up to nearly $2 trillion. Credit crisis has been costly, but perhaps it has not yet erased all the savings achieved so far.

As usual, the story is a lot more complex than this simple calculation shows. Dismantling the securitization machine would not help us to prevent future bubbles and crashes. Instead, it is likely to make us poorer. We should focus instead on how to limit our systemic risks.

A Year after Bear Died

Bear Stearns died a year ago. Its stock was sold for $2 per share to JP Morgan. That was a shocking offer to market. We though the wires had a typo, they probably meant $20. A lot of wealth got destroyed, a lot of people got unemployed at that time. It was tragic!

Looking back, it seems naïve of us that we got shocked by what happened then. We lived up seeing much worse. Bear employees and stockholders actually got lucky in some ironic sense. $2 stock offer was raised to $10. Much better than $0 it would have reached otherwise. Most of the employees who were laid off found jobs elsewhere. The job market was not as bad at the time.

Still, it was not that hard to predict that economy and markets would get worse. Despite the two month long rally following the Bear’s fold into JP Morgan’s hands, this was a clear sign that the credit crisis had gotten to a new stage. Already, the government was the only hope, only reasonable solution out of this crisis. Over those months, many financial firms resorted to distortions to minimize their damage. History showed that a financial crisis was always costly, very costly.

Yet, things could have turned around by the end of 2008. It was not certain but possible. Paulson committed huge mistakes in September 2008. His mistakes were avoidable. He saved Bear Stearns because it was a system-critical player and then he let Lehman go bankrupt. Even though Lehman was several times more inter-connected or critical than Bear, he did not seem to have the wisdom or courage to do the same for Lehman.

I was surprised at that time not so much by Lehman’s bankruptcy but by the inaction of the US government. They did not seem to have a plan ready to put to action. Everyone knew after Bear Stearns, Lehman was next. Fed knew it! They asked even hedge funds about their exposure to Bear Stearns *and* Lehman Brothers in March 2008. Six months is plenty time to develop contingency plans.

Perhaps, you could say, the government was indeed prepared. Let Lehman die, but save AIG. Either there was no plan, no preparation or this indeed was the plan itself.

Credit Investments of Warren Buffett

What is the greatest investor doing in this credit crisis? He is spending most of his cash on corporate debt or preferred stock. We don’t see him making large company acquisitions as of yet. He sees value in debt not equity.

Table: Warren Buffett’s investments in credit since the start of credit crisis:

This speaks volumes!

Buffett is not alone in going where the value is. The world’s richest person by some ranking, Carlos Slim, also likes this idea. He recently lent $250 M convertible debt to New York Times company at 14%. Another shrewd investor, Carl Icahn, was reported to be collecting the debt of MGM at 38 cents on dollar in March ‘09.

CDS of GE

GE is in the news once again! This venerable company, probably the best industrial firm in the world, is being hugely discounted in the market. Its equity below $7, market cap below $70 B, credit markets are pricing its bonds and CDS as if it might go bankrupt.

Two questions can be asked: Does GE have a real risk of bankruptcy? If not, what is the reason for such violent reaction in the markets?


First, does GE have real bankruptcy risk? GE’s CDS has quickly shot up, going beyond the levels where Bear Stearns and Lehman were trading before their collapse. To emphasize again, these are the danger territory that not even Bear and Lehman has reached before their end. CDS markets indeed imply that there is a large probability of bankruptcy for GE. The cumulative probability of default over 5 years is about 50% as implied by the market spreads of roughly 1000 bps for 5 year CDS (arbitrarily assuming 40% recovery rate given default). At 20% upfront with 5% coupon over 5 years, protection buyers are paying almost half the notional that they will recover in case of default:


This reaction in CDS markets have been followed by equity and equity derivatives markets.


Taking Feb 11 as base (where there has not been a particular news for GE), we can observe the reaction in equity markets has been less moderate and lagging compared to CDS markets:



Part of the reason for more violent reasons in CDS markets arise from the put option like characteristics of CDS. Even taking into account the optionality of CDS, it is hard to reconcile the CDS market’s violent reaction with the fundamentals.

GE’s problems stem from the concerns with its looming rating downgrade. GE’s financial arm, GE Capital, is a AAA-rated company which by no means deserves that rating in this new environment where the counterparty risk is paramount. GE Capital has benefited from its AAA rating in past, but that might be now its Achille’s Heel. It might have to post collateral for certain transactions in which it has provided credit protection.

It is well known by now that the rating agencies will soon cut the rating of GE. The only question is how harsh that downgrade would be. Depending on the extent of this downgrade, GE could require substantial cash to post as collateral for its certain derivatives trades. Like a lot of highly rated companies, the financial engineers in GE were no doubt monetizing this credit rating during the credit bubble. Credit intermediation trades, credit guarantees, providing long-term liquidity with short-term funding, these are just some of the well known examples. It is not inconceivable that GE might experience what happened to AIG, a spiral down scenario where downgrade begets further credit weakness.

Is bankruptcy of a GE Capital a reasonable scenario then? Not really! Yes, GE may have to post a large amount of collateral (or come up with cash) in case of a harsh rating downgrade. But, here is the rub! GE Capital benefits from the easy funding facilities provided by US Treasury. This company is selling bonds, guaranteed by FDIC, under the TGLP program with very low spreads to US Treasuries. These bonds have a large demand, because they are effectively guaranteed by US government.

In order to GE to go bankrupt, the government will have to reverse its programs and ideology of supporting the large financial institutions. This is of course possible, but not very likely in the near term. Just as the AIG analogy is harmful to GE, the explicit and implict support of government is beneficial for GE.

Sharp moves in GE’s CDS are by and large involves the dysfunctionality of CDS markets. Fundamental credit indicators do not indicate a bankruptcy at all (adjusting the parameters for the new financial world). GE’s CDS has climbed up in huge amount mostly because of the lack of liquidity in CDS markets. CDS market has lost substantial liquidity. There are no natural sellers of protection. As soon as a bad news hits, it creates a demand for protection which finds a supply only at extraordinarily high rates. In old times, there would be many people who would be willing to take this risk. They could then lay off this risk in different ways. For example, dealers could have sold bespoke CDOs where the risk is sold off. No more of that in this environment. As a result, those couple of players who really need to get protection, will pay for it through the nose.

Then, we have the second question: If the bankruptcy risk is not concordant with the CDS levels, what is driving the valuation? GE’s issue is not bankruptcy but the potential for further equity dilution as it may have to raise capital. The government may not be so generous in future as now and GE Capital may have to come to market. This is somewhat similar to what happened to Citigroup and other financial institutions. Such a huge entity cannot be supported endlessly by the government. GE Capital will have to stand on its merits. Time might be a healer, but the market is not giving the company the benefit of doubt.

CDS Volumes May be a Factor of Instability

CDSs have become notorious for many reasons. Some of these perhaps arise from a misunderstanding of this opaque market. Yet, there are also very legitimate concerns. CDS trading has not mitigated, on contrary, helped proliferate the counterparty risk. Those with insider information have preferred this market as they don’t leave easily spotted tracks there. Moving CDS trading to exchanges would no doubt help in such issues. In an important issue, however, moving to exchange would not make a difference. That concerns the size or volume of CDS trading. For the sake of financial system, we all must seriously consider if there should be limits on the size of this market.


The notional amounts involved in this market are huge. While these notional amounts vary from week to week, they are in the order of tens of trillion dollars, or roughly about the same size of stock markets. As CDSs are fixed income instruments, the risk amounts are indeed a fraction of these huge notionals involved. Furthermore, most of these instruments are actually traded between dealers, with customers making only much less. Only about 10% of the notionals will remain after the netting of all contracts.

Still, CDS trading notionals are large amounts. In some cases of CDS trading, the notional amounts traded are larger than the total debt outstanding of the reference entity. For some, this is a spot-on example of tail wagging the dog. Without a direct link between the volume of the referenced debt and the corresponding CDS amount traded, it may be hard to have a stable credit market. Yves at Naked capitalism proposes that the scale of the risks involved in CDS market may be too large even for an exchange.

I list below some of the most striking discrepancies between the CDS volume and the total debt for non-financial US corporations:

Table: CDS volumes May be Multiples of Outstanding Debt

Notional CDS volumes above come from DTCC and are as of mid January 2009. I have computed the total debt outstanding by adding the loan and bond amount issued by each corporation and its subsidiaries. The total outstanding debt would be smaller in many cases, but I preferred to stay with the most conservative assumptions.

The most egregious case is for GAP. For this company with $550 million debt, market participants have bets on specific CDS contracts with total notional of $23 Billion., 42 times larger. Even after netting of all offsetting CDS, the net notional is 3 times greater than the total debt.

Moreover, the example above understates the traded volume of CDS. GAP is a component of several credit indices. The above table doesn’t include what is bet on this company using such indices. The amounts could easily double if we could take the index trading into account. Second, there could be additional amount of custom contracts not captured in the DTCC data. Third, there could be swaps on the bank loans outstanding that will amplify the bets on the debt of the company.

As a counterargument to this concern, one can note that the notional amounts involved in CDS are not very relevant. Perhaps, in the initial stages of this market, CDS was mainly a risk mitigating contract, removing the default risk. Now, CDSs are used mainly to express opinions on the direction of credit worthiness of an entity. Outright bets or hedges are not intended to be kept until their maturity, instead they would be unwound and sometimes rolled every quarter. Futures market may provide a useful analogy. Just as the futures market, most participants will close out or roll their existing contracts before the expiry date. Traded volumes can thus be greater than what is deliverable physically.

CDS market takes physical delivery issue one step further. Even after a credit event, most contracts can be settled by cash as opposed to physical delivery that was used when the market was new. Removing the physical delivery rule may be where the CDS got unhinged from the underlying debt markets. IRA’s Chris Whalen considers this one of the main weaknesses of this market:

"The basic model for a CDS contract does not really fit the needs of investors or the real economy, who are in the most simplistic terms looking for a practical way to hedge an illiquid corporate bond. Unfortunately, most corporate bonds cannot be borrowed in the securities lending market, WHICH MEANS THAT THERE IS NO TRUE CASH BASIS FOR SINGLE NAME CDS. Faced with this issue, the happy squirrels at ISDA came up with cash settlement as a way to ensure the astronomical growth of CDS - never realizing that in so doing, they were also magnifying the overall level of risk in the global financial system many times over and above the actual "basis," represented by the bonds specified in each CDS contract."


Such rationale has also been the behind the proposed plans by New York State Insurance Department Superintendent Eric Dinallo concerning the restrictions on naked CDS trades, i.e. those CDS trades not used for hedging cash positions.

Disconnect between the CDS and underlying debt may have serious implications for corporations. Their access to capital is determined in part by this market. As the trading volume and liquidity in the corporate CDS market much better than the corresponding cash debt market, CDS markets are usually the first place to reflect the pertinent news on credit prices. They affect not only credit, but also equity and volatility markets as well. We have already seen in last few months salient instances of risk perceptions to be self-fulfilling. The concerns of investors induce CDS buying and as the resulting CDS spread move further, these trigger more CDS buying. A positive feedback mechanism!

Can such instability occur at systemic level?... 1987’s stockmarket crash involved the use of index futures by portfolio insurance followers as the market dropped. In that case, the coupling and feedback mechanism between the cash market and futures market triggered a sudden drop in markets. Without rules and limitations on the size of CDS trading, it is hard to eliminate such a possibility.

Credit Spreads versus Equity Dividend Yields

Credit and equity divide the claims on the same assets of a corporation. Capital structure arbitrage tries to benefit from any mispricing between them. Yet, as often is applied, the term “arbitrage” is misnomer. Credit and equity are different markets. Mispricings discovered will often involve taking some sort of risk. Still, a careful observer can draw insights from the relative pricing between them, and might even be able to profit from them.

One of the simpler ways of checking pricing information between credit and equity markets involves mapping credit spreads versus dividend yields. For instance, the following graph shows the 5y CDS spreads versus indicated dividend yields of select US corporations as of mid January 2009:

Should we expect any simple relationship between the credit spreads and dividend yields? Let’s look at the market by segmenting into three parts: high spread, medium spread and low spread companies:

• On the high spread segment of the market, dividend yields may range from very high to zero. These are all stressed companies running a serious risk that their dividend would have to be cut. We would expect companies whose debt trades at high spreads would have high dividend yields reflecting this risk (like CIT, XL), unless they have already taken action to cut it down.

• On the low spread segment of the market, is natural to see low dividend yields. Companies placed here, by and large, are those that do not have serious issues with any debt payment. Stability in their profits would usually allow them to plow their earnings back to business. This is especially so if they are in the growth markets as opposed to mature markets. A new crop of corporations in this segment involve those whose credits are backstopped by government.

• In the medium spread markets, we would observe companies exposed to an event risk, such as an impending merger or acquisition (for example CTL, DOW). These can have high or low dividend yield depending on the expected structure after the event.

Still, the relationships between the credit spreads and dividend yields seem to be more complex than the basic relationships described above. There are some companies with relatively high dividend yields in all segments even without obvious event risks. MO and CBS are some examples that stand out.

Is there any inconsistency between the equity and credit markets for these companies? If we exclude the special cases, such as those corporations with unstable dividend payment policy or near term liquidity issues, would we be left with hard to explain cases, or potential mispricings?

First thing to note is that each example may have very good reasons to have its credit and equity trade at those prices. Capital structure is often complex. Existence of bank loans above the senior debt and any subordinated debt and preferreds between the senior debt and common equity would affect the pricing relative to equity. Credit claims can range from debt due immediately to those with 50 years maturity and those with equity warrants or hybrids or converts. All of this means, a sweeping generalization would be misleading.

It may yet be possible though to discover some “information” reflected in one market but not the other. A proof of that would use a trading strategy based on historical data. The first step in creating such a trading strategy would involve taking the CDS spreads, normalizing them by index spreads and mapping them against the dividend yield and equity prices. Second step is the creation of a mechanical trading rule that focuses on outliers. Third step involves the elimination of false signals by excluding the special cases.

Before looking at the results of such a trading strategy, let’s review whether such a trading strategy would make sense fundamentally. The most basic relationship between the credit and equity is through Merton’s capital structure model. The key model of this insight is that the equity owners have a call option on the balance sheet of a company. Equity is then priced a call option on the value of the assets of the company.

where A signifies the value of assets, F signifies the face value of equity and σ_A is the volatility of assets. Asset volatility is again determined through the Black-Scholes formula from the equity volatility:

d2 above is a measure of distance-to-default, and N(-d2) signifies the risk-neutral probability of default. Thus, the part of credit spreads due to expected loss can be derived in the Merton framework. This model yields a simple inverse relation between the credit spreads and indicated dividend yields.

There are, of course, many many issues with this model which is based on extremely simplified assumptions. (For example, insolvency does not necessarily lead to default in real world). This basic model has been enhanced in many directions by researchers. Some involve the use of non-zero barriers and pricing of knock-and-out options. Some involve stochastic volatility. Some focus on distance to default and a more elaborate representation of capital structure (such as in KMV models).

However, the basic Merton model is still useful in the sense that, it reminds us the following:

Asset/equity volatility is a key parameter in the debt-equity relationship.

Going back to our question on whether there is any mispricing, it is clear that we cannot ignore the equity volatility if we try to exploit information from one market in another.

Thus, we must modify the above trading strategy by incorporating the equity volatility as a parameter for the beginnings of a model...

Buyers and Sellers of US Treasuries

Foreign Central Banks have sold a very large amount of US Treasury securities in November 2008. The selling of US Treasuries by foreigners has reached $23 Billion in that month, the largest outflow since January '78:

Who are selling and who are buying the US Treasuries? US TIC data shows the trends:

In the last 12 months,

• China has bought over $200 B of US Treasuries, increasing its inventory by 45%.
• UK has also bought over $200 B; this represents a doubling of their inventory.
• Inventory has also doubled in off-shore banking centers that added $100B.
• Oil exporters and Russia account for another $100 B increase.
• Japan has kept its inventory nearly the same over this period.

Those who reduced their inventory in November 2008 are: Japan, Brazil, Russia, Korea, Luxembourg. I guess Luxembourg’s action must be at least partly related to unwind of structured transactions that include Treasuries in them.


Pricing Review:
The following chart shows the yield of 10Y US Treasury note. After spending most of 2008 in the range of 3.5 – 4.0% yield, its yield suddenly started dropping in mid November 2008, pretty much straight until 2% at the year end.


The foreign central bank purchases cannot explain the move of this magnitude! It is not a result of diminishing supply either as late 2008 was a period of increasing supply of US Treasury debt. This rally stemmed from the purchases by private sector: Banks were trying to reshape their balance sheet by moving from risky assets to US government debt (I would never call them risk-free!) in order to present a soothing balance sheet by the year-end.


Future Trend:
Given the fast drop in oil prices, we should not expect more buying of US Treasuries by oil exporters that we saw in 2007. Don’t know why UK doubled their US Treasury inventory, but it shouldn’t be easy for them to repeat this as they are a spender. Amid the increasing signs of discomfort with US debt by Chinese officials, Chinese buying of US Treasuries is already slowing down. Yves at naked capitalism discusses this issue pointing out that Chinese kept buying US Treasuries to prevent RMB appreciating too much against $. (Append: Later, though, she concurs with Brad Setser who claims China will keep buying US Treasuries as the reduction in their exports will be smaller than the reductions in their imports).

It is hard to see the foreigners buying more US Treasuries in 2009. Of course, we have a big buyer right here in US that plans to buy whatever US Treasury must sell. Fed!