PrefBlog

An interesting paper by Bernadette A. Minton & René Stulz & Rohan Williamson is How Much Do Banks Use Credit Derivatives to Hedge Loans?:

Before the credit crisis that started in mid-2007, it was generally believed by top regulators that credit derivatives make banks sounder. In this paper, we investigate the validity of this view. We examine the use of credit derivatives by US bank holding companies with assets in excess of one billion dollars from 1999 to 2005. Using the Federal Reserve Bank of Chicago Bank Holding Company Database, we find that in 2005 the gross notional amount of credit derivatives held by banks exceeds the amount of loans on their books. Only 23 large banks out of 395 use credit derivatives and most of their derivatives positions are held for dealer activities rather than for hedging of loans. The net notional amount of credit derivatives used for hedging of loans in 2005 represents less than 2% of the total notional amount of credit derivatives held by banks and less than 2% of their loans. We conclude that the use of credit derivatives by banks to hedge loans is limited because of adverse selection and moral hazard problems and because of the inability of banks to use hedge accounting when hedging with credit derivatives. Our evidence raises important questions about the extent to which the use of credit derivatives makes banks sounder.
…
Our evidence helps understand why the use of credit derivatives for hedging is limited. First, the market for credit derivatives is the most liquid for investment grade corporations and for countries. As a result, use of credit derivatives is going to be more intense for firms that have exposures to such credits, which we find to be the case. Second, for non-investment grade corporates, the market for credit derivatives is less liquid. Further, private information is more important for banks for loans to such corporates. As a result, hedging will be more expensive and banks will hedge such loans less. Using disclosures of banks, we find that banks that report hedging across credit ratings hedge relatively more credits that are less risky, which is consistent with our prediction. Finally, hedge accounting cannot typically be used for credit derivatives.
…
For 2005, we show that the total credit protection bought and sold by banks is roughly $5.5 trillion. In comparison, the net protection bought, which is a measure of hedging of credit risks, is roughly $0.5 trillion, or less than 10% of the overall credit derivatives gross positions of banks. While, the net protection bought is small compared to the loans of the banks that have positions in credit derivatives, the gross position of these banks is large compared to the loans they write. Consequently, since credit derivatives are used only to a limited extent to hedge loans, they can only make banks and the financial system sounder if they create few risks for banks when the banks take positions in them for other reasons than to hedge loans. Contrary to the optimistic view of regulators before 2007, the subprime crisis has shown that the dealer positions of banks in credit derivatives have substantial risks.

I’m not sure if it makes much sense for banks to systematically hedge their loan exposures through CDS. It seems to me that the whole point of being a bank in the first place is to hedge your issuer-specific risk through diversification and excess margin.

This is interesting in light of the CIT affair. When CIT drew down its credit lines – as reported on PrefBlog on March 20, 2008, CDS spreads spiked up to 27% up front and 500bp p.a. instantly – I recall, but cannot substantiate the existence of, rumours to the effect that this was due to bank buying.

This would make more sense, since CIT was drawing on credit lines with a pre-arranged spread. The Boston Fed looked at this issue. It is also my understanding that a lot of new lines are being negotiated as a spread off of CDS levels, which strikes me as a much better use of CDS by banks.

I can only hope the authors will re-examine the issue as data from the Credit Crunch trickles in!

James Hamilton of Econbrowser has announced a new paper he has co-authored: The Market-Perceived Monetary Policy Rule:

We introduce a novel method for estimating a monetary policy rule using macroeconomic news. Market forecasts of both economic conditions and monetary policy are affected by news, and our estimation links the two effects. This enables us to estimate directly the policy rule agents use to form their expectations, and in so doing flexibly capture the particular dynamics of policy response. We find evidence that between 1994 and 2007 the market-perceived Federal Reserve policy rule changed: the output response vanished, and the inflation response path became more gradual but larger in long-run magnitude. In a standard model we show that output smoothing caused by a larger inflation response magnitude is offset by the more measured pace of response. Our response coeffcient estimates are robust to measurement and theoretical issues with both potential output and the inflation target.
…
Our baseline results for the 1994-2007 sample suggest the market perceives that the Federal Reserve gradually responds to inflation and real activity. Similar to previous literature working on post-Volcker data, we find the Federal Reserve follows the Taylor Principle, a greater than one-for-one response to inflation. We also find evidence that the market-perceived monetary policy rule changed over our sample. During the 1990s market-perceived policy responded robustly to output and quickly to inflation; during the 2000s market-perceived policy doesn’t respond to output and responds at a more measured pace to inflation, though its long-run inflation response is greater than before. We quantify the importance of the inflation response path and long-run magnitude in a standard model, and find that raising the long-run magnitude is effective at lowering inflation volatility while making the path more gradual is counterproductive. Our baseline results were found to be robust to alternative possible specifications.

The Bank of Canada Review, Summer 2009 has been released, with the articles:

• Collateral Management in the LVTS by Canadian Financial Institutions
• The Complexities of Financial Risk Management and Systemic Risks
• The Changing Pace of Labour Reallocation in Canada: Causes and Consequences
• BoC-GEM: Modelling the World Economy

The question of Collateral Management, addressed in the first article, sprang to prominence at the height of the Credit Crunch in 4Q08, when the “eligibility premium” – the yield differential between two securities identical in all respects except that one was eligible to be used as collateral with the central bank, the other not – exploded from its normally immeasurably small levels.

The LVTS [Large Value Transfer System] is a real-time, electronic wire transfer system that processes large-value, time-critical payments quickly and continuously throughout the day. Participants in the LVTS use claims on the Bank of Canada to settle net payment obligations. To secure the payments that are sent through the LVTS, collateral is required.
…
The Bank originally accepted only Government of Canada (GoC) securities as collateral, but since it expanded the list in November 2001 to include a larger variety of securities (e.g., municipal securities and commercial paper), pools of collateral pledged by individual FIs to the LVTS have diversified significantly. Thus, while GoC-issued securities constituted about 55 per cent of the discounted value of securities pledged in 2002, they made up less than 30 per cent in early 2007
…
The results of this study are important for policymakers such as the Bank of Canada, which is concerned both about the effi cient functioning of fixedincome markets and about the credit risk it ultimately bears in insuring LVTS settlement. Given these new insights into the behaviour of FIs, future changes in collateral policies, in particular those regarding the eligibility of assets as collateral, can be designed more effectively.

Ongoing monitoring of and research into collateral management practices is required to keep abreast of
the changing behaviours at fi nancial institutions and within an evolving financial environment. Future
research will examine collateral management in more detail, with a particular focus on changes resulting from the recent fi nancial crisis and the ensuing increase in Government of Canada debt issuance.

The second article is also of interest, although I suspect that it is merely another volley in the battle for the BoC to extend its influence to macro-prudential regulation and frustrate the designs of OSFI upon the turf:

Banking theory has made very limited inroads into the theory and practice of risk management, where modelling has been dominated by the frictionless, efficient-market model masquerading under the title of financial engineering.
…
For example, in 1998, Salomon Brothers (as related in Bookstaber 2007, Chapter 5) were using a model of the yield curve, the so-called two-plus model (two random factors plus a constant—with the constant signalling shifts in Federal Reserve policy). The model had worked well to produce a steady stream of arbitrage profits over several years. In 1998, these profits changed to a stream of losses as the fixed-income arbitrage group struggled with what seemed to be a change in the underlying model. It seemed that another random factor had appeared, leaving the group holding residual risks, which were causing large losses. The risk manager struggled to help the group, but in the end, it was shut down. The exit had to be disguised and undertaken over several weeks, since Salomon’s large positions in the market were affecting bond liquidity and could entice arbitrageurs to exploit the company. The worst-case scenario would have occurred if Salomon’s sales had driven down prices, leading other traders to dump bonds and driving prices even further down, thus exacerbating Salomon’s losses. Bookstaber argues that this exit by Salomon’s large bond-arbitrage group made the market less liquid and increased the difficulties faced by Long-Term Capital Management (LTCM) later in the year, when its bond-arbitrage position became untenable after the Russian bond default (another unmodelled risk).
…
In all the above models, three major risks stem from model misspecifi cation through either: (i) choosing the
wrong number of random factors; (ii) inappropriate random factor distributions (e.g., normal, symmetric
distributions rather than skewed distributions), and/or (iii) using poor parameter estimates for the coefficients or factor loadings on risky factors. These risks should be tested regularly by back-testing the models llooking for systematic deviations from the model using actual data), and checking the history of trades and the profi t/loss outcomes on exposures. Because all models are merely approximations, losses and profits on exposures should be expected. In a well specified and calibrated model, however, the history of profits and losses will expose biases. Any detected biases should be examined, and appropriate action taken. Although this is easy to state as a general principle, in reality, the management and estimation of risks is far from perfect, especially in periods of high volatility, where correlations can change rapidly.