Data processing: financial – business practice – management – or co – Automated electrical financial or business practice or... – Finance
Reexamination Certificate
1998-05-28
2001-08-21
Hafiz, Tariq R. (Department: 2163)
Data processing: financial, business practice, management, or co
Automated electrical financial or business practice or...
Finance
Reexamination Certificate
active
06278981
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to the field of data processing, and in particular to a computer-implemented method and apparatus for compressing a portfolio of financial instruments to enable, for example, more efficient risk management processing than is otherwise achievable with an uncompressed portfolio.
Risk management is a critical task for any manager of a portfolio of market instruments, and accurate and efficient risk measurement is at the core of any sound enterprise-wide risk management strategy. Given the relatively-complex mathematical calculations necessary to accurately measure risk, financial institutions generally use some form of computer-implemented “risk management engine.” As explained below, however, existing risk management engines may be insufficient to adequately deal with the large, complex portfolios maintained by many financial institutions.
It is not unusual for large and medium-sized financial institutions, such as banks or insurance companies, to require a risk management engine that allows the computation of daily Value-at-Risk (VaR) estimates of an entire portfolio, which may contain several hundred thousand positions, including substantial volumes of complex derivative products such as swaps, caps and floors, swaptions, mortgage-backed securities, and so on. Moreover, these several hundred thousand positions may have to be evaluated over hundreds or even thousands of different scenarios. To further complicate the task, these financial institutions may require decision support tools for managers and traders that allow performance of inter-day calculations in near-real time.
In general, financial institutions are required to measure their overall risks for regulatory purposes and as a basis to manage their capital more efficiently. While the former has been driving the development of risk oversight programs in financial institutions worldwide in the last few years, the latter provides a high value-added to those willing to make the investment. Traditionally, portfolio managers have been using standard deviation and variance to measure their portfolio risk. This practice is based on modern portfolio theory, as described in, for example, Harry Markowitz, Portfolio Selection, The Journal of Finance, vol. 7, no. 1 (1952), and W. F. Sharpe, Capital Asset Prices: A Theory of Market Equilibrium Under Conditions of Risk, The Journal of Finance, vol. 19, no. 3 (1964). However, in the last decade, both regulators and businesses have embraced more general (and perhaps more sophisticated) measures such as Value-at-Risk. VaR gives the maximum level of losses that a portfolio could incur, over some predetermined period of time, with a high degree of confidence. For regulatory purposes, for example, the time period may be set to 10 days, and the one-sided confidence interval to 99%. See, e.g., Planned Supplement to the Capital Accord to Incorporate Market Risks, Basle Committee on Banking Supervision, Bank of International Settlements, Basle, No. 16 (April 1995). Although VaR can be expressed as a multiple of the portfolio standard deviation in some simple cases, such as when portfolios are normally distributed, this generally is not the case.
There are different methods available to estimate VaR, depending on the assumptions one is willing to make with respect to the possible future market moves and the complexity of the portfolio. Such methods are described generally in RiskMetrics™ Technical Document, Morgan Guarantee Trust Company Global Research (4th ed. 1996), and Phillipe Jorion,
Value at Risk. The New Benchmark for Controlling Derivatives Risk
(Irwin Professional Publishing 1997). The most generally-applicable method is based on simulation, either historical or so-called “Monte Carlo” simulation. In particular, some simulation may be unavoidable to get an accurate picture of risk when a portfolio contains substantial positions in instruments with optionality, such as options, convertible bonds, mortgages and loans with embedded options. However, given the complexity and computational requirements of known simulation methods, users must trade accuracy for price, time and ease of implementation. Moreover, full simulation of very large and complex portfolios, such as those encountered in many financial institutions today, may not be achievable in a reasonable time period even with top-of-the-line computers. For example, a VaR estimate of a large, complex portfolio over several thousand Monte Carlo scenarios could easily take several hours, if not days, for a top-of-the-line work station. Indeed, even the simple task of loading and storing large portfolios can be onerous and time consuming.
In an effort to address the practical problems associated with risk measurement for large and/or complex portfolios, it is known to adopt an approach in which a subject portfolio (also called the “target” portfolio) is first divided into a “linear” subportfolio and a “non-linear” subportfolio. The former would contain all of the instruments having little or no optionality, while the latter would contain all of the options. In a typical institution, the linear portfolio might comprise 70-95% of the total portfolio positions. However, given their nature, the risks embedded in option positions may be substantial. The next step in such an approach is to measure the risk of these subportfolios separately. For the linear subportfolio, one could apply, for example, a “delta-normal methodology” such as that described in the above-cited RiskMetrics™ Technical Document. By assuming linearity of the subportfolio and normal distributions, this analytical method has moderate computational requirements. For the options, some basic, perhaps limited, simulation can be applied. Finally, an estimate of the risk of the target portfolio is taken as the sum of the individual subportfolio risks.
A significant problem with this approach, however, is presented by the last step. To illustrate, consider a simple example where a trader sells a call option on a given bond and immediately buys a hedge on the underlying bond. Although the bond clearly reduces the portfolio's risk, the above-described methodology would indicate that the VaR of the portfolio has increased (and in fact almost doubled). In general, a mix of methodologies may grossly overestimate VaR since it fails to account for the main principles of risk management: hedging and diversification. This may result in undesirable penalties for good risk management policies.
In view of the shortcomings with known approaches for risk management of large and/or complex portfolios, including but not limited to the shortcomings discussed above, it is apparent that there is a need for a computer-implemented process that is capable of representing such portfolios in a compact way, and that achieves such compression (e.g., loads instruments, generates cashflows, compresses, etc.) quickly and efficiently. Likewise, in contrast to the division approach discussed above, there is a need for a single methodology that enables measurement of risk across an entire portfolio. Such a single methodology should offer sufficient computational efficiency to permit accurate risk measurement to be completed in a reasonable time period regardless of the size and/or complexity of the target portfolio. Embodiments of the present invention satisfy these and other needs.
SUMMARY OF THE INVENTION
The present invention is generally directed at providing improved tools for risk management of large and/or complex portfolios of financial instruments. In accordance with particular embodiments of the invention, as described herein, a “compressed portfolio” is generated for a given target portfolio of financial instruments. In general, the compressed portfolio is a relatively smaller and/or simpler portfolio that closely mimics the behavior of the target portfolio, but that requires orders of magnitude less computer memory to store and orders of magnitude less computational time to value. Thus, the compressed portfolio can be used,
Dembo Ron Samuel
Kreinin Alexander Yacov
Rosen Dan
Algorithmics International Corporation
Hafiz Tariq R.
Kenyon & Kenyon
Meinecke-Diaz Susanna
LandOfFree
Computer-implemented method and apparatus for portfolio... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Computer-implemented method and apparatus for portfolio..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Computer-implemented method and apparatus for portfolio... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2487949