Monday, September 24, 2012

Bringing Academic Rigor to Government Bond Ratings

Since peaking in July, yields on Italian and Spanish long-term bonds have dropped by about 150 basis points. While the headlines attribute this sharp adjustment to the availability of a new ECB bond buying program, neither country has expressed an intention to use it, so perhaps the headline writers are missing something. Looking back, our July 23rd research note concluded that the spread between Italian and German bond yields was excessive given the low probability of an Italian sovereign default.

A Problem of Information

Market volatility is compounded when market participants lack the information necessary to appropriately value investment securities. When rumors and announcements penetrate the information vacuum, investors often overreact. If this vacuum is instead filled with high-quality information and analysis, volatility and mispricing diminish.

Equity market investors can obtain research and analysis from numerous banks and brokerage firms. While much of this research suffers from conflicts of interest, it also contains large volumes of fact and analysis that investors find useful.

For government bonds, the traditional source of analysis has been the major credit rating agencies. Unfortunately, these firms have faced widespread criticism in recent years, leaving their reputations in tatters and their guidance in doubt. Although the highest profile failures have been in structured finance, critics have also questioned whether incumbent rating agencies have sufficient staffing and resources, adequate procedures, and the intellectual capability to meaningfully gauge sovereign default risk.

Potential Ways Forward – A Call to Academics

A number of parties have recommended alternatives to the incumbent rating agencies, including at least four not-for-profit initiatives offering sovereign ratings.

To have a beneficial impact, a non-profit rating agency will have to gain credibility with investors. After all, if ratings don’t guide investment decisions, what value do they have? A not-for-profit credit rating agency can gain credibility by having a strong methodology and solid institutional support. Both of these factors can be advanced by the academic community. Economists, political scientists, statisticians and financial engineers associated with a major university could generate the kind of quality, branded sovereign research that would command the attention of investors.

In hopes of focusing more academic attention on government credit risk, I have contributed an article to the new issue of Economics Journal Watch – a peer-reviewed economics journal freely available online. The article describes problems with the rating agency model, surveys some of the previous literature on government default probability modeling, and offers a research agenda as well as one possible solution.

Academics can deliver the intellectual rigor missing from both status quo rating analysis and some of the alternatives we have been seeing. The thorough data collection procedures, advanced modeling techniques, and peer review practices employed by social scientists can raise the level of sovereign risk analysis.

Wednesday, September 19, 2012

One Bond, One Price

The contemporaneous "scandals" recently covered in the media - the alleged manipulation of LIBOR and the possibility that JPMorgan used its VAR model to help disguise the riskiness of its portfolio - return us directly to the very shortcomings that led to, or exacerbated, our financial crisis. 

They provide ready examples of the frailty of our financial controls, and acute reminders of what can happen when financial institutions have the incentive, and capacity, to massage financial data and disclosures. 

It ought not to surprise anybody that, when trying to maximize their gain under a poorly-designed incentive structure, financial market professionals might choose to manipulate valuations or massage disclosures to their advantage. Would not most players maneuver to their benefit? 

The problem here, of course, is their capacity to massage important data on a large scale. Back in 1994, Kidder, Peabody's Joseph Jett was purportedly able to exploit an anomaly in Kidder's accounting system, booking substantial profits in the absence of any genuinely profitable trades. 

Almost twenty years after the "Kidder Scandal," our system remains open to abuse. 

The lack of transparency in our reporting, and the ability to finesse asset valuations, create an environment in which the value of a bank's assets quickly becomes whatever the bank wants it to be. Banks and funds need little incentive to inflate their asset prices: higher asset valuations typically lead to improved performance, which enables banks or funds to raise more capital at reduced rates. Imagine if mortgage borrowers could value their own houses and have their mortgage rates reduced based on their own inflated appraisals! 

It is time to realize that the current system does not work - it cannot. We can no longer rely on an institution to accurately evaluate its own assets, while being knowledgeable of the many and material conflicts it faces. We also cannot blindly trust the auditing firms to catch the cover-ups. Their success rate has been too low. 

One advantage of the ratings-based system - a system from which we're energetically moving away - is that it provided a single rating for each bond. It is true that banks may reserve different amounts of capital against each rated bond, but at least we knew that it was always one bond, one Moody's rating.  

Right now, however, there is virtually no pricing control - no consistent mechanism for the valuation of complex and illiquid transactions. One bond can have many prices, depending on who provides the price. We see regular examples of traders who mark their own portfolios, while earning bonuses and promotions based on their self-constructed performance. We often find that two similar banks, supervised by the same regulator, can hold the same security at materially different prices. JPMorgan itself was recently reported to have held the same bond in two different divisions at different prices. Banks can "shop" for opportunistic evaluations, just as they could actively seek the highest ratings, which precipitated a lowering of ratings standards. How does this all promote shareholder confidence, or cross-institutional comparability? 

Goldman Sachs, Wachovia, Bank of America, Citigroup and others have recently been accused of mispricing securities in one way or another. Traders at UBS, Deutsche Bank, Credit Suisse, RBS have been criticized for mismarking bonds. Hedge funds and asset managers have been accused of overvaluing their funds' positions and the audits of Deloitte, PwC and KPMG have all come under scrutiny for their failure to capture their clients' misrepresentations and securities mispricings. 

The list keeps growing. In last week’s case filing In re Lehman Brothers Holdings Inc., et al v. JPMorgan Chase Bank NA, the Lehman plaintiffs argue that the JPMorgan entities’ derivatives claims are inflated as a result of procedures including “the inconsistent valuation of trades to the JPMorgan Entities’ advantage.” 

While the problem of inconsistent, utility-maximizing pricing is overwhelming, the solution is simple. It only requires the political power for its implementation. All financial institutions should be forced to carry the same bond at the same price. Any digression from market pricings should be accompanied by appropriate disclosure that describes the reason for, and magnitude of, the deviation. Furthermore, any internally-marked positions should be open to public scrutiny especially for those institutions considered "Systemically Important Financial Institutions" (or SIFIs). What is there to hide, anyway? 

Supervisory authorities would have only to decide on the framework for deriving the ultimate price. There are several options available, not the least of which is to follow the precedent set by the National Association of Insurance Commissioners. The NAIC selected PIMCO and Blackrock Solutions, respectively, to help it determine risk-based capital requirements for insurers’ holdings of residential and commercial mortgage-backed securities. Rather than calculating capital reserves, one could implement an algorithm that produces a price. Another approach would be to create a central pricing repository which would accept a range of prices submitted by regulator-approved price providers. The final price could then be determined by applying an averaging process, for example, or by taking the median of the submissions. 

Pricing transparency, and a forum for the provision of feedback to the extent assets are being mispriced, would enhance investor confidence in the reliability of banks' balance sheets. Confidence and transparency, of course, lead to market liquidity. After the crisis of confidence we have had, and the associated liquidity freeze, wouldn't that be nice? 

Additional Resources