Agenda

Abstracts

Jason Halliwell

Jason holds a commerce/law degree with first class honours in commerce from UQ, and a grad dip in maths and finance from UTS. This has been followed up with 6 years of industry experience working on quantitative tactical asset allocation modelling, the first three years at Westpac Investment Management, and more recently as research manager for asset allocation at GMO Australia.

Finding Value: modelling in the world of flawed accounting

In the market hysteria leading up to the lofty peaks of what we now know to be the tech bubble, the market was awash with new paradigm euphoria. We now know that rules haven't been re-written, but there is still a lot of misinformation about where the long-term equilibrium really is. This presentation looks at some very widely used and accepted equity valuation models, and exposes their theoretical and empirical short-fallings. These shortfallings can be traced to systemic failures in our accounting system, and can largely be blamed on the complete inability to deal with the distorting effects of inflation. Only after adjusting for these failures are we able to gain a true insight into where the bottom will be - and why todays bargain hunters may be in for a rude shock.

Johnathan Calvert

Jonathan Calvert is a partner at Grantham, Mayo, van Otterloo, a Boston-based money management firm with 25 billion USD in assets under management. Prior to joining GMO, he developed knowledge-based systems in the telecommunications industry. In his ten years at GMO Mr. Calvert has been involved in the development of the quantitative research environment and research and portfolio management for both the US and International quantitative portfolios. His current focus is on trading research for the efficient implementation of the quant portfolios. Mr. Calvert received a Bachelor of Mathematics in Computer Science from the University of Waterloo and is a Chartered Financial Analyst.

Optimal Portfolio Execution

Price Impact is generally a function of the amount traded. To reduce the price impact of portfolio trades, traders can spread trades over multiple time periods. Increasing the duration of a trade, however, can increase the timing cost and risk of the trade list.

In this talk, the author will discuss the use of stochastic dynamic programming to model trading sequences based on stock-specific return and cost predictions over a finite time horizon. An optimal set of trades is determined using integer programming optimization to minimize the total cost of implementing portfolio trades subject to portfolio cash flow constraints. This analysis is extended to include fair allocation of trades across multiple accounts. Results for simulated and actual trade lists are analyzed.

Volf Frishling

Graduated with First Class Honours from the University of Tbilisi in the Republic of Georgia in 1972. Upon graduation joined the Academy of Sciences of Georgia, while simultaneously doing part-time post-graduate studies.

Immigrated to Australia in 1981 and obtained Ph D Degree from the Flinders University of South Australia in 1982, specialising in Stochastic Differential Equations and Optimisation of Controlled Markov Processes.

Since arrival to Australia worked in various private and government organisations prior joining the State Bank of Victoria in 1990. In 1993 was appointed Head of the Quantitative Analysis Group at the Commonwealth Bank of Australia, Financial Markets with responsibilities for pricing all derivative products traded by the CBA. In 2001 was appointed Chief Manager, Quantitative Research - Risk in the Institutional Banking division of CBA.

In 2001 was appointed Adjunct Professor at University of Technology, Sydney.

Published more than 10 papers, several latest in financial applications.

Generation of Co-integrated paths for Credit Exposure Measurement

The CBA Credit Risk Engine (CRE) uses full Monte Carlo simulation to generate future market scenarios. Standard methods generate such paths using correlated Brownian motions. One of the criteria for the quality of the model should be the correct behaviour of some simulated ensembles, eg all interest rates within the same economy should move more or less in parallel.

This is not necessarily the case if standard methods are used, for example it has been observed that for many simulated Australian and US interest rate path ensembles:

• rates diverge significantly,
• rates are in incorrect relationship, for example the 5 year swap rate is higher than the 10 year and remains so for long time.

The impact of this on the performance of the CRE is twofold:

• operational - CRE generates a large number of errors because of the failure of the yield curve building procedure due to inconsistent inputs, and
• quantitative - incorrect portfolio exposures are generated.

This paper proposes, a new, more realistic model based on the concept of co-intergation. Paths generated by this model exhibit much more believable behaviour and this results in reduced exposures.

John Okunev and Derek White

An Alternative Approach to Measuring Value at Risk of Hedge Funds

This work analyses the risk characteristics of hedge funds. Specifically, the returns of the hedge fund styles and its constituents are regressed (mapped) onto subsets of 31 actual indices, the directional components of the indices, and option-like payoff patterns derived from the 31 indices. Once these mappings are complete, the value-at-risk for the hedge funds and its style components is estimated using simulations of actual historical returns and the mappings onto various indices. We find that value at risk of hedge funds is much greater using this approach when compared to the mean variance approach.

I. Outline

The appropriate methodology by which to evaluate the risk exposure to investing with hedge funds has, as of late, received increasing attention in the academic literature. In general, the limitations for traditional asset classes to adequately encompass the risk characteristics embodied within hedge fund returns is now widely recognised and researchers have been searching for alternative methodologies for estimating risk exposure. Hedge funds, unlike traditionally managed funds, are much more free to initiate long or short positions across a wide array of asset classes and markets at will. In addition, hedge funds are much more likely to utilise derivatives or futures contracts than are more conventional funds. Even without initiating derivatives positions, these hedge funds may utilise various dynamic trading strategies that cause their after-fee returns to exhibit various option-like payoff patterns - the protective put providing one simple example. Moreover, these trading strategies may even be utilised to manage exposure to the second moment through either increasing or decreasing risk contingent upon prior relative performance. Even without considering the trading strategies of hedge funds, the fee structure within the hedge fund industry itself produces option-like post-fee returns to investors. High watermark provisions and fees contingent on absolute performance impose an implicit short call position to the investors of the hedge fund. In short, strictly linear factor models may fail to adequately model the true risk exposures to after-fee hedge fund returns.

In recognition of these limitations, academic research has begun to make use of contingent claims analysis to model the underlying risk to hedge fund returns. The traditional method to evaluate fund managers is to regress the fund's historical returns on a set of benchmarks. The slope coefficients from the regression provide the benchmark-related exposures and the intercept ("alpha") gives an estimate of performance after controlling for factor sensitivities. This methodology can be traced back to Jensen (1968). Unfortunately, this type of analysis is sensitive to nonlinear relationships between the fund and factor returns and may lead to erroneous inferences regarding relative performance. Grinblatt and Titman (1989) have shown that a fund manager can generate positive "risk-adjusted returns" by simply selling call options on the underlying assets of the portfolio. A standard means to control for option-like return features is to add a nonlinear function of factor returns as independent regressors. In fact, this methodology has been used in Treynor and Mazuy (1966) and Henriksson and Merton (1981).

Glosten and Jagannathan (1994) develop a theoretical model to use benchmark-style indices that have embedded option-like features. This methodology has been used to evaluate the performance and risk characteristics of hedge funds. Agarwal and Naik (2000) demonstrate that well over half the variability to hedge fund returns may be explained by long / short positions in options on various factors. Mitchell and Pulvino (2001) show the returns to risk arbitrage are similar to those obtained from selling uncovered index put options. Fung and Hsieh (2001) argue that trend-following hedge fund returns are best modelled by using the payoffs to a lookback straddle. The purpose of this work is to examine the risk characteristics for hedge funds and its style constituents. We confine our analysis of the style components to those that have continuous returns data for this period which include: MultiProcess Group, Arbitrage Fixed Income, Arbitrage Total, Relative Value MultiProcess, Relative Value Total, Security Selection Long-Bias, Security Selection No-Bias, and Security Selection Total. Initially, our analysis will identify the underlying risk factors for hedge funds. As part of this analysis, we will examine the stability of the factors across the seven year time period of the study. Once we have completed the factor analysis, we will estimate the value-at-risk for the hedge fund and for each of its style. We find that using this methodology value at risk is substantially higher when compared to the traditional mean variance approach.

Alan Scowcroft

A new global country-sector model

The increase in industrial concentration with the growth of global stocks has led to a biased sector composition in many smaller markets. This makes it very difficult to separate the impact of domestic market movements from global industry effects and as a consequence is an important issue for both portfolio construction and risk measurement.

Since an increasing number of markets are dominated by a small group of stocks or sectors there is also a significant danger that apparently high levels of correlation are in fact spurious. This is particularly so when a few stocks can account for a large proportion of the market capitalisation of a single country, region or global sector.

This talk will consider an extension to the model of Heston and Rouwenhorst, simultaneously constructing sector neutral local maket indices and country neutral global sector indices. However, the beta's of local industries are free to vary within a set of identifying restrictions. The estimation of the model will be discussed and contrasted to alternative approaches and the results illustrated at both stock and portfolio level.