Papers by Clemens Glaffig
This article proposes a method to optimize portfolios of hedge funds, taking investor preferences... more This article proposes a method to optimize portfolios of hedge funds, taking investor preferences as the starting point to define an objective function that will be flexible enough to include general investor preferences. In particular, we can include path-and marketdependent objectives. The article also develops a method to produce a forward-looking data set on which the optimization can be based. Our method is also particularly apt for a case when a portfolio of hedge funds is used as an overlay. We compare this approach to classical optimizations on empirical data to highlight the effects of the additional degrees of freedom we have included.
AESTIMATIO : the IEB International Journal of Finance, 2011
The rise in popularity of benchmark free and complex trading strategies throughout the last decad... more The rise in popularity of benchmark free and complex trading strategies throughout the last decade has made available a large variety of risk and performance profiles. As a consequence, to account for their complex performance characteristics, a lot of effort has been devoted to classify and value the performance of these strategies by the alterations of previous-or innovative measures. However, as most measures are often still simple path-and context independent statistics, most often the information provided proves inadequate to separate performance characteristics-as evidenced by the latest crisis. This paper provides a methodology that integrates the clustering and performance measurement of trading strategies in a context and preference based environment. It decomposes preferred performance characteristics into fragments of context dependent behaviour for clustering purposes. It subsequently aggregates these fragments of performance characteristics into a performance measure. The methodology allows for consideration of path dependencies. Two applications, in the clustering of hedge fund styles and the ordering of alternative equity strategies are given. A further application in the statistical replication of trading strategies is highlighted.
Social Science Research Network, Jan 23, 2006
This article proposes a method to optimize portfolios of hedge funds, taking investor preferences... more This article proposes a method to optimize portfolios of hedge funds, taking investor preferences as the starting point to define an objective function that will be flexible enough to include general investor preferences. In particular, we can include path-and marketdependent objectives. The article also develops a method to produce a forward-looking data set on which the optimization can be based. Our method is also particularly apt for a case when a portfolio of hedge funds is used as an overlay. We compare this approach to classical optimizations on empirical data to highlight the effects of the additional degrees of freedom we have included.
Social Science Research Network, 2018
In this treatise, we describe some of the basic and common deep learning architectures amenable t... more In this treatise, we describe some of the basic and common deep learning architectures amenable to objectives in finance and give heuristic arguments as to which of those seems preferable for applications in financial trading. A specific deep architecture, allowing information cascades and different feature weights for different market environments, is described. Applicable objectives and input feature engineering for this architecture are discussed.
Palgrave Macmillan UK eBooks, 2012
Social Science Research Network, 2010
The rise in popularity of benchmark free and complex trading strategies throughout the last decad... more The rise in popularity of benchmark free and complex trading strategies throughout the last decade has made available a large variety of risk and performance profiles. As a consequence, to account for their complex performance characteristics, a lot of effort has been devoted to classify and value the performance of these strategies by the alterations of previous � or innovative measures. However, as most measures are often still simple path � and context independent statistics, most often the information provided proves inadequate to separate performance characteristics � as evidenced by the latest crisis. This paper provides a methodology that integrates the clustering and performance measurement of trading strategies in a context and preference based environment. It decomposes preferred performance characteristics into fragments of context dependent behaviour for clustering purposes. It subsequently aggregates these fragments of performance characteristics into a performance measure. The methodology allows for consideration of path dependencies. Two applications, in the clustering of hedge fund styles and the ordering of alternative equity strategies are given. A further application in the statistical replication of trading strategies is highlighted.
Journal of Functional Analysis, Sep 1, 1990
Smoothness of the integrated density of states, k(E), of random Schrodinger operators, i.i.d. and... more Smoothness of the integrated density of states, k(E), of random Schrodinger operators, i.i.d. and non-i.i.d. cases, on a discrete strip lattice is investigated. It is proven that k(E) is C" if only the potentials on the top surface of the strip have distributions with compactly supported densities in some fractional Sobolev space. The C=-result for the case of the Anderson model, i.e., all potentials having a distribution with compactly supported density in some Sobolev space, is also recovered.
Journal of Statistical Physics, Apr 1, 1987
It is shown that the probability distribution for the infinite-volume, free-boundary-condition Is... more It is shown that the probability distribution for the infinite-volume, free-boundary-condition Ising ferromagnet on the Bethe lattice under zero external field is infinitely divisible with respect to the group operation of pointwise multiplication of spin variables.
Journal of Statistical Physics, 1987
It is shown that the probability distribution for the infinite-volume, free-boundary-condition Is... more It is shown that the probability distribution for the infinite-volume, free-boundary-condition Ising ferromagnet on the Bethe lattice under zero external field is infinitely divisible with respect to the group operation of pointwise multiplication of spin variables.
The rise in popularity of benchmark free and complex trading strategies throughout the last decad... more The rise in popularity of benchmark free and complex trading strategies throughout the last decade has made available a large variety of risk and performance profiles. As a consequence, to account for their complex performance characteristics, a lot of effort has been devoted to classify and value the performance of these strategies by the alterations of previous- or innovative measures. However, as most measures are often still simple path- and context independent statistics, most often the information provided proves inadequate to separate performance characteristics- as evidenced by the latest crisis. This paper provides a methodology that integrates the clustering and performance measurement of trading strategies in a context and preference based environment. It decomposes preferred performance characteristics into fragments of context dependent behaviour for clustering purposes. It subsequently aggregates these fragments of performance characteristics into a performance measure...
SSRN Electronic Journal
In this treatise, we describe some of the basic and common deep learning architectures amenable t... more In this treatise, we describe some of the basic and common deep learning architectures amenable to objectives in finance and give heuristic arguments as to which of those seems preferable for applications in financial trading. A specific deep architecture, allowing information cascades and different feature weights for different market environments, is described. Applicable objectives and input feature engineering for this architecture are discussed.
We investigate smoothness properties of the integrated density of states ( ids ) for random Schro... more We investigate smoothness properties of the integrated density of states ( ids ) for random Schrodinger operators on a multidimensional strip lattice, where only the potentials on the "top surface" of this lattice have a distribution with some regularity. We view the eigenvalue equation on the strip as the action of an abstract group on some homogeneous space, from where we derive a representation of the ids in terms of a distinguished measure on that homogeneous space. This representation allows us to conclude that using minimal smoothness of the potential distribution on the "top surface", combined with a negative moment condition for the distribution of all other potentials, is enough to obtain smoothness of the ids. This includes the original Anderson model. We also discuss cases, where the distribution of the potentials below the "top surface" is Bernoulli, satisfying this negative moment condition.
Ssrn Electronic Journal, Dec 22, 2010
The rise in popularity of benchmark free and complex trading strategies throughout the last decad... more The rise in popularity of benchmark free and complex trading strategies throughout the last decade has made available a large variety of risk and performance profiles. As a consequence, to account for their complex performance characteristics, a lot of effort has been devoted to classify and value the performance of these strategies by the alterations of previous � or innovative measures. However, as most measures are often still simple path � and context independent statistics, most often the information provided proves inadequate to separate performance characteristics � as evidenced by the latest crisis. This paper provides a methodology that integrates the clustering and performance measurement of trading strategies in a context and preference based environment. It decomposes preferred performance characteristics into fragments of context dependent behaviour for clustering purposes. It subsequently aggregates these fragments of performance characteristics into a performance measure. The methodology allows for consideration of path dependencies. Two applications, in the clustering of hedge fund styles and the ordering of alternative equity strategies are given. A further application in the statistical replication of trading strategies is highlighted.
This article proposes a method to optimize portfolios of hedge funds, taking investor preferences... more This article proposes a method to optimize portfolios of hedge funds, taking investor preferences as the starting point to define an objective function that will be flexible enough to include general investor preferences. In particular, we can include path-and marketdependent objectives. The article also develops a method to produce a forward-looking data set on which the optimization can be based. Our method is also particularly apt for a case when a portfolio of hedge funds is used as an overlay. We compare this approach to classical optimizations on empirical data to highlight the effects of the additional degrees of freedom we have included.
Ssrn Electronic Journal, Apr 3, 2011
We propose a new hybrid hedge fund replication technique, which combines aspects of portfolio con... more We propose a new hybrid hedge fund replication technique, which combines aspects of portfolio construction from factor based replication with an innovative version of distributional replication. It uses a parameterized replicating strategy for which we match a version of state contingent integral stochastic dominance. The dominance will be over a set of distributions reflecting preferred, state contingent distributional performance characteristics, granting insight into return features to arbitrary fine detail. It emphasizes the replication of desired aspects rather than the replication of the performance path. A further application of this approach is to replicate certain features of a target fund and at the same time dominate on less desired aspect. Before the new approach to replication is presented, a brief recollection of the evolution, the various different approaches and some of the pitfalls of hedge fund replication are highlighted.
Ssrn Electronic Journal, Aug 20, 2004
The article investigates the use of adaptive learning algorithms in constructing dynamic portfoli... more The article investigates the use of adaptive learning algorithms in constructing dynamic portfolios replicating the return characteristics of a given hedge fund. The emphasis is on out of sample conditional predictive capabilites as necessary to serve as a valuable risk management tool, rather than simply explaining hedge fund behaviour over an in sample period. The algorithms learn dynamic trading rules and strategies along with which factors to base those on, within an integrated learning mechanism. It thus generalizes previous approaches by exploring a wide class of nonlinear and dynamic trading strategies to participate in explaining and predicting hedge fund behaviour. The conditional predictive capabilities of the algorithms can specifically be employed to quantify future fund behaviour. It will be useful in constructing quantitative risk measures for individual hedge funds. The article also provides some empirical data for out of sample behaviour of this method.
Aestimatio the Ieb International Journal of Finance, 2011
The rise in popularity of benchmark free and complex trading strategies throughout the last decad... more The rise in popularity of benchmark free and complex trading strategies throughout the last decade has made available a large variety of risk and performance profiles. As a consequence, to account for their complex performance characteristics, a lot of effort has been devoted to classify and value the performance of these strategies by the alterations of previous-or innovative measures. However, as most measures are often still simple path-and context independent statistics, most often the information provided proves inadequate to separate performance characteristics-as evidenced by the latest crisis. This paper provides a methodology that integrates the clustering and performance measurement of trading strategies in a context and preference based environment. It decomposes preferred performance characteristics into fragments of context dependent behaviour for clustering purposes. It subsequently aggregates these fragments of performance characteristics into a performance measure. The methodology allows for consideration of path dependencies. Two applications, in the clustering of hedge fund styles and the ordering of alternative equity strategies are given. A further application in the statistical replication of trading strategies is highlighted.
Uploads
Papers by Clemens Glaffig