• Selected
• Entire List

Abstract:
From the risk management's perspective, one of the main differences between asset management companies and banks concerns the investment horizon: typically, asset managers have longer investment horizons. We compare different ways to deal with medium/long horizons, when the aim is to calculate absolute or relative VaR using a historical simulation approach and its variations, like bootstrapping procedures. We use several indices to test the accuracy of the different methods analysed. We find these methodologies:

- can provide satisfactory assessments of tactical risk;

- can inform portfolio managers of changes in market risk;

- are also promising for strategic risk analysis.

Asset management, risk,historical simulation, bootstrapping, value at risk, relative value at risk

Abstract:
The article concerns the differences between the meaning of risk management in a bank and in an asset management company, and then it illustrates the solution found to the challenge of building up a risk management system in an Italian medium size company.

We describe the specific needs of an asset management company, in terms of proper financial modelling, time and resources constraints, given that an investment company:

- manages "third party funds" (eventual losses are not its liabilities);

- has a relatively long investment horizon, if compared to other market participants;

- often monitors relative risk, rather than absolute risk.

As models for estimating portfolio risk on medium-long horizons are crucial, and the related theoretical problems are absolutely not trivial, we briefly describe the model we use. It is based on bootstrapping, that allows us to calculate several risk measures for a large number of portfolios. It allows us to simulate medium-term to long-term financial scenarios, without imposing any particular probability function, taking into account heteroscedasticity, serial correlation, and any existing market views.

We shortly describe the implementation, with some details on the database and the calculation engine (both available on the market). This kind of system is more straightforward and cheaper to implement than traditional, bank-oriented risk systems: it took less than one year to implement the whole tool, with limited costs (basically some cheap licences and the work of two quants).

We hope that our experience could be of help to other asset management companies.

Risk management, asset management, implementation, bootstrap, historical simulation

Abstract:
We present a multi-period risk model to measure portfolio risk that integrates market risk, credit risk and, in a simplified way, liquidity risk. Thus, it overcomes the major limitation currently shared by many risk models that are unable to give a complete picture of all portfolio risks according to a single, coherent framework. The model is based on the Filtered Bootstrap approach; hence, it captures conditional heteroskedasticity, serial correlation and non-normality in the risk factors, that is, most of the features of observed financial time series. Being a simulation risk model, it copes in a natural way with derivatives as it allows the full valuation of the probability density function of the contracts. In addition, it is a suitable and flexible way to generate future scenarios on mediumterm horizons, so this model is particularly appropriate for asset management companies.