Michael A H Dempster is Managing Director of Cambridge Systems Associates Limited and Professor Emeritus, Centre for Financial Research, Department of Pure Mathematics and Statistics, University of Cambridge. 

As the founder of the Centre for Financial Research in Cambridge University and having been a consultant to many global financial institutions, Michael discusses his most current research on derivatives.


Current Research of M A H Dempster

University of Cambridge & Cambridge Systems Associates

The past four decades have seen the exponential growth of derivative markets which at 2014 mid-year stood at a total mark-to-market value of about US$17.5T and a notional value of over US$ 691T -- i.e. about 10 times global GDP! -- according to the latest BIS figures. At this point the proportion of derivatives traded on exchanges reached only a little less than 11%, the remainder being bespoke over the counter (OTC) products. Two aspects of my research on OTC derivatives address the questions of the title: the first is theoretical and the second practical.

The aim of the theoretical work with Ezequiel Antar is to provide a rigorous but realistic theory of how the risk reduction benefits of trading innovative financial products are realized and distributed amongst market participants. Of necessity this theory must be one of incomplete markets in which new products traded cover risks previously un-priced. The dynamic mathematical representation of trader behaviour must include their portfolio allocations of traded instruments -- new and old -- involving their attitudes to risk embodied in suitable risk measures. Dynamic trading equilibria are derived together with the design of new securities. To this end we have developed a theory of backward stochastic equations in discrete time, since the known theory in continuous time is inadequate to characterize trading equilibria in the general case. While the new theory involves serious mathematics, the concepts and results can be illlustrated with a simple intuitive computational example.

The practical work is related to financial litigation regarding OTC derivative products. Over the past seven years, Cambridge Systems Associates has been engaged by clients involved in litigation with financial institutions over structured option, swap, bond, FX and equity financial products. CSA has been asked to apply its expertise to the evaluation of a wide variety of contracts struck between various counterparties in the period from 2004 to date. We have also recently collaborated with Bloomberg News in understanding some high profile cases in the Eurozone. These usually involve global financial institutions and counterparties in various jurisdictions: Germany, India, Italy, Ireland, Monte Carlo, Switzerland, Japan, the UK and the US to date. The clients are firms, second tier banks, local authorities and high net worth individuals in these jurisdictions. Notional principal of the deals range from a few, to hundreds of millions of pounds. Our reports, based on mathematical models and  historical market data, give fair market prices for the contracts and present additional results in the form of projected net cash flows between the client and the bank. Their net present value distributions provide clear visualizations of the relative risks of a deal for its counterparties and an intuitive translation of the mathematical concepts involved.  

This work has given me a clear insight into the true nature of the OTC derivative markets. Last year at Global Derivatives Bruno Dupire stated that "we all know that derivatives are about money up front and regulatory arbitrage" and then discussed a structured fixed floating amortized swap whose current MTM was over 5 times its notional which he called the "worst trade of all time". But we have valued a twice restructured amortized CMS10 memory swap whose current MTM was nearly 7 times its notional! In both cases the counterparties were governmental, common targets of the issuers.

Some stylized features of the hundreds of structured OTC derivatives we have valued to date are the following. Each deal represents a play by the issuing bank that exploits their superior knowledge of possible future market evolution relative to the client's. Issuers are often the client's commercial bank and the term sheets/contracts usually bear a feminine bank signature. Often the bank requires the deal as a condition of a loan, refinancing or bond flotation. Each deal is structured to have the enticement of a short term client "sweetener" which can sometimes be very subtle. Enticement can even be buried in a programme of successive similarly structured deals which only in the later stages become egregious -- playing the "fish". Due to severe asymmetry of information the client is in no position to understand their risk relative to the bank's, which is often extreme. When a deal begins to go wrong for the client, the bank offers to postpone the agony by restructuring the deal to one even worse. In summary, products are invariably mispriced in favour of the bank at inception and get worse over time. For the issuer it's like going to the track having fixed the horse race. You are not absolutely guaranteed to win, but you surely have an edge on the punters!

One might ask why do clients ever sign such structured OTC contracts? Unfortunately, rational theory involving random search by rational risk neutral investors is of little help. Darrell Duffie does however say that "premia are higher when [issuers] are harder to find, when [buyers] have less bargaining power, when the fraction of qualified [clients] is smaller, and when risk aversion, volatility, or hedging demand is larger". Recently, Thorsten Hens  showed using behavioural theory that structured investment products offer no gains after fees to rational investors; they appear to do so only if the client mis-estimates outcomes by overweighting favourable outcomes relative to unfavourable ones. Is this a theory of gullibility? Why clients actually sign such OTC contracts can be ranked on a scale from the honest weak to the business and politically powerful: namely, desperation (Detroit), coercion (US and UK SMEs), trust (Mittelstand), gullibility (Landesbanken, Italian cities) and complicity (Greek and Italian governments, Monte dei Paschi di Sienna). The US Dodd-Frank act has clear rules enforcing client duty of care and separating advice and trading with clients, but driven by Wall Street interests it is currently under Congressional attack. The rest of the world has a very long way to go to return derivative markets to their early proper role in reducing financial risk, and progress is glacial!