WO2004042532A2

WO2004042532A2 - Financial planning constraints for portfolio analysis

Info

Publication number: WO2004042532A2
Application number: PCT/US2003/035004
Authority: WO
Inventors: Brian Hollander; Curtis Burton Downing
Original assignee: Mellon Investors, Llc
Priority date: 2002-11-01
Filing date: 2003-11-03
Publication date: 2004-05-21
Also published as: WO2004042532A3; AU2003291695A1; AU2003291695A8

Abstract

Systems and methods for the automated generation of an asset allocation and automated analysis of an investment portfolio. Information regarding an individual investor is preferably used to generate an individualized asset allocation that takes into account the financial goals of the investor. The present system utilizes this asset allocation or an asset allocation generated by any other means to analyze and optimize a portfolio. The present system preferably evaluates the investment products held by an investor and all available investment products to generate a list of investment products to be pruned, augmented, or in which the investor should newly invest. This analysis preferably proceeds using explicit rules for pruning, augmenting, or newlyinvesting in specific investment products. The rules may be configurable so that different approaches to portfolio analysis and optimization may be employed easily.

Description

FINANCIAL PLANNING CONSTRAINTS FOR PORTFOLIO

ANALYSIS

CLAIM OF PRIORITY

This application claims the benefit under 35 U.S.C. 119(e) of U.S. Provisional Application No. 60/423,229 filed November 1, 2002 and claims continuation-in-part status under 35 U.S.C. 120 of co-pending U.S. Utility Application No. 10/230,872 filed August 22, 2002.

BACKGROUND OF THE INVENTION

FIELD OF THE INVENTION

The present invention relates to systems and methods for portfolio analysis and for the generation of recommendations for investments within an investment portfolio for an investor. The present invention also preferably provides for the collection of user-specific information for use in generating an asset allocation when recommending investments within an investment portfolio.

DESCRIPTION OF THE BACKGROUND

Financial planning is the process of examining a broad array of information about an investor's financial life to determine whether specific financial goals can be achieved. Further, financial planning is the basis on which one may create a broad investment strategy that may be used in generating an optimized asset allocation in cash, bonds, and stocks. Generally, financial planning approaches take into account the cash flows of an investor and the timing and magnitude of financial goals.

Multiple factors are traditionally considered when developing an investment portfolio. One of the first questions to address is how should financial assets be allocated among such categories as stocks, bonds, and cash. In addition, among each of those categories an investor must determine in which sub-category (e.g., particular types of stocks) and which investment product (e.g., a specific mutual fund) he or she will invest in order to optimize the portfolio. When making all of these decisions the investor or his financial advisor may consider additional factors that are particular to the individual investor such as risk tolerance, net worth, age, and financial goals.

Traditional techniques for creating an optimized portfolio do not apply an asset allocation constraint based on a thorough examination of the investor's financial planning needs. For example, a financial institution may recommend a pre-set distribution of financial assets for investors based only on their net worth and age, rather than based upon any specific, future financial goals or based solely on a view of near or long term market performance. It would be preferable to generate an asset allocation in light of the specific financial planning needs of an investor so as to provide recommendations that reflect a truly personalized determination of their financial goals.

However, this type of analysis must be performed on a case-by-case basis and can be very time consuming for both the investor and an investment professional. To take into account the age, present net worth, and timing and magnitude of future significant events to generate an asset allocation for an investor requires extensive interaction between an investor and a financial advisor. This process is inefficient and laborious. There has been a long standing need within the financial industry for an automated portfolio analysis tool that provides for the generation of an optimized portfolio across all of an investor's accounts that fully reflects an asset allocation based on personalized financial planning analysis.

Once an asset allocation has been generated, portfolio optimization can occur. Recommendations for pruning of investments, augmenting of investments, and for new investments are developed in light of the asset allocation. This process may also be influenced by the investment preferences of the investor. For example, an investor may have a strong aversion to particular types of investments or strong preferences for others. An automated portfolio analysis tool would preferably incorporate these pieces of information while at the same time applying automated rules that allow for the assessment and optimization of financial portfolios.

SUMMARY OF THE INVENTION

In accordance with the present invention, there are provided systems and methods for the automated generation of an asset allocation and automated analysis of an investment portfolio for the purpose of pruning, augmenting, and adding individual securities. Initially, information is preferably collected from an investor. This information includes the investor's present distribution of assets in investment products, saving rate, risk tolerance, age, the timing and cost of financially- significant future events, and the financial goals. By using this information collected about the investor, an individualized asset allocation will preferably be generated.

The present system utilizes this asset allocation, or an asset allocation generated by any other means, to analyze and optimize a portfolio. To perform this task, the present system preferably has access to information about all available investment products. The products are characterized within the present system by their attributes such as expected and past return on investment, investment style, asset class, expense ratio costs, as well as other relevant factors. Explicit rules are preferably used to evaluate investment products in light of these attributes and to generate recommendations for pruning, augmenting, or newly-investing in specific investment products. The rules may be configurable so that different approaches to portfolio analysis and optimization may be employed easily.

DETAILED DESCRIPTION OF THE INVENTION

The present invention preferably allows for the automated generation of an asset allocation based on all relevant financial information about an investor and the automated analysis and optimization of an investment portfolio. The present system preferably incorporates information regarding an investor including, but not limited to, total net worth, the timing and magnitude of future financial requirements and goals, risk tolerances, and net worth of the investor to arrive at an indicated asset allocation. This asset allocation is preferably used as a constraint in the analysis of the distribution of an investor's investments when creating an optimized portfolio. Analysis and optimization of an investor's portfolio will result in recommendations that take the form of pruning (i.e., reducing or eliminating) investment in a security, augmenting investments in a security, or newly investing in a security. Preferably, the present system provides for the automatic suggestion of these actions so that portfolio analysis and recommendation generation occur without requirements for intervention by an individual such as a financial advisor.

Detailed information is collected for each security that is available to an investor. This helps define the universe of products in which an investor would or should invest. These, in part, determine the part of the pruning and augmenting parameters. These parameters may be further defined by a financial advisor and/or investor. Databases that contain this information allow the present portfolio of an investor to be defined clearly for the application of various rules. Once these parameters are represented in this way, an existing portfolio is analyzed to generate recommendations for securities that investor should prune, augment, or invest in light of both an asset allocation and a sub-asset class allocation methodology. Pruning of a user's portfolio may be considered screening (in part or all) investment products that are deemed unacceptable. Augmenting of a user's portfolio may be considered adding newly-recommended financial products that are selected to fill deficiencies in the portfolio's given the objective of creating an optimized portfolio. Optimization of a portfolio determines precisely how much money should be invested in each investment product in each investment category in each account in the investor's portfolio.

The present system may be directly employed by individual investors or investment companies. The term 'users' will be employed herein to refer to both individual investors and companies. The phrase "investment firm" will be used herein to refer generally to any individual or entity that provides financial advice to a user through operation of the present system. The investment firm may configure multiple aspects of the present invention. These include the universe of investment categories into which investment products may be classified and how investments are categorized. The investment may also configure the system by adjusting the rules for classifying users or the rules for determining whether to prune or recommend a stock or fund.

Initially, detailed information is obtained from or about a user. This information may be acquired through an automated interview process or through the completion of a questionnaire and/or through the transfer into the system of aggregated data about the investor. During this process, the total amount that is invested/available to invest, age of the investor, and the risk tolerance of that investor is obtained. In addition, the particular securities in which the investor is presently invested and the total amount of money invested in those securities is provided by the investor. The timing and magnitude of any future financial goals is also ascertained. For example, the investor would indicate that he or she is planning on paying for a college education for a child in six years or if he or she is planning on making a major purchase in the next five years.

Some of this information is preferably employed to generate an asset allocation. An asset allocation is defined as the target percentage distribution of an investor's investments among the three main investment classes, i.e. stocks, bonds, and cash. Such data as age, savings rate, revenue sources, income, and financial goals of an investor directly impact the manner in which a determination is made about the percentage of the financial assets that should be invested in each of these three classes. Preferably, the core system is used to create asset allocation as a constraint for the portfolio analysis system. For example, if the asset allocation dictates that 30% of an investor's assets be in stocks, 50% in bonds, and 20% in cash, the portfolio analysis portion of the present invention will preferentially recommend an optimized portfolio that satisfies those constraints, hi generating an asset allocation, the present system may consider such issues as the need for immediately-available cash to accomplish a particular financial goal or the timing of investment product maturation and financial goals.

In addition to personal financial and biographical information, investors may specify particular securities that may not be pruned, augmented, or added to their portfolio. For example, an investor may, for sentimental reasons, wish to retain a particular stock or bond or alternatively, wish to avoid investments in specific financial products. Each piece of information is represented within the system so that it may be used during the development of recommendations for portfolio modification.

An asset allocation that is generated in other manners may also be employed by the present system. For example, if an investment firm has a standardized asset allocation for investors of a certain age and net worth, the present system may directly use this asset class constraint within the present system to perform portfolio analysis and optimization.

In order to appropriately perform an analysis of the user's portfolio and generate an asset allocation, the present system preferably categorizes users according to certain pre-established rules. Each user of the present system may be categorized according to any user data that were collected from the user. The user categories table defines user categories and provides the present system with rules to categorize users.

A user preferably should meet every criterion for a particular User Category in order to be identified as belonging to that user category. If a user does not meet the criteria for any category, that user is categorized as belonging to the default user category. This categorization will preferably influence the generation of an asset allocation and portfolio analysis and optimization.

To perform portfolio analysis and optimization, the present system preferably represents information regarding various investment products. Typically, these investment products are publicly available, such as stocks, mutual funds, and treasury bills. However, the present invention is also useful in the analysis of privately-held investment products. The data collected about an investment product allows the present system to generate recommendations for the pruning and augmentation of an investment portfolio.

In both analyzing the products currently held by an investor and generating recommendations for investment strategies, the present system preferably assigns specific characteristics to every security available to an investor. These characteristics are used properly to categorize each investment product in an sub- asset category, e.g. large cap, small cap, foreign, international, short term bond, etc. Generally, the categories are established according to guidelines used by the investment product information provider such as S&P, Lipper, Morningstar, etc. Once received the information about each security is preferably stored in a database for use in later analysis and is periodically updated when the data provider provides updated information.

Each investment product recognized by the present system belongs to one Investment Category. The present system may define the entire universe of investment products that an investment firm offers. Note that this is distinct from defining which investment categories the present system will recommend. The present system may divide investment products first into top level asset classes (i.e., stocks, bonds, or cash) and then into investment categories (e.g. large cap value mutual fund, mid cap foreign mutual fund, individual stocks, taxable bonds, tax- exempt bonds, etc.). The present system may use as many investment categories and sub-categories as is appropriate given the investment firm's rules to establish an optimized portfolio up to all investment accounts, for the portfolio that is being analyzed. An example of the manner in which the present system may categorize investment products is present in the following table.

For the categorization of mutual funds, the present system may rely upon the peer group code that is provided by a data provider for each mutual fund. For categorizing bonds, the average maturity of a bond is preferably used. Short term bonds may be defined as having average maturities of less than three years and long term bond funds as having average maturities of greater than then years.

Returns-based style analysis is preferably performed on each security to refine the categorization process. Each index of an investment product that is used in the style analysis is assigned an expected return based on available long-term data. The security's return is then computed by taking the appropriate weighted average of the investment styles attributed to it. Mutual funds' expected returns are adjusted downward by their expense ratios. In addition, tax-exempt bond funds are scaled by a factor that represents the current ratio of municipal bond yields to taxable bond yields. Expected returns are used to categorized the style of a security and categorize the security within the databases of the present invention.

The style analysis also preferably takes into account dividends when analyzing the style of an investment product. The expected annual dividend is computed by annualizing the total growth of a security due to dividends or other distributions over the last five years. The dividends are not generally considered to be returns above and beyond the expected return computed as described above. They merely represent the taxable portion of that return.

As a further example of categorization of securities, a U.S. individual equity security is defined as a small cap if its market value is less than, for example, $500 million, large cap if their market value is greater than $5 billion, and mid cap if it is between those two values. Within these capitalization categories, equities may be categorized as growth, blend, or value depending on the book-to-price ratio. The top fraction (e.g., a third) of products in book-to-price ratio is considered 'value' and a middle fraction is considered 'blend' or 'value' depending on the growth characteristics of that fund. U.S. equities with unreported market capitalization or book-to-price ratio are categorized by way of style analysis. The domestic equity ι index most highly correlated with the stock's or fund's returns determines the investment category for the stock or fund.

Because of sporadic fundamental data, all American Deposit Receipts (ADR's) are classified according to style analysis. The foreign index to which an ADR's returns are most highly correlated preferably determines the investment category for the ADR.

The system preferably updates data in the databases regarding the securities on a monthly basis to reflect new market information about the securities. The monthly securities data processing script populates the Calculated Data table after the raw data from data vendors are imported. The Calculated Data table contains the investment category, expected return, and expected dividend for each security. It may also be customized to contain other fields, which may then be used as attributes in the Pruning Rules table, as described below.

Since the securities data processing script is customized, the method of categorizing securities and computing securities' expected returns is up to the investment firm operating this present system. The default is to use returns-based style analysis, which is configured with the Style Indices table. A security's expected return is estimated to be the weighted average of the expected returns of the style attributions' for that security and may be represented in the following manner.

Each security is categorized in the above described manner and the information is collected in a database. This process helps define the universe of products in which an investor has an investment or could invest. These databases allow the present portfolio of an investor to be defined clearly for the application of various rules as described hereinbelow. Once the portfolio of an investor is represented in this way, rules are applied to the portfolio to generate recommendations for securities that investor should prune, augment, or invest in light of an asset allocation. Pruning of a user's portfolio may be considered screening mutual funds or stocks thatiare deemed unacceptable. Augmenting of a user's portfolio is adding newly recommended mutual funds that are selected to fill deficiencies in the portfolio's market coverage. Optimization of a portfolio determines precisely how much money should be invested in each stock or found in each specific account..

Many of the runtime tables include a description of a criterion that is used to define a rule that the portfolio construction algorithm follows. A criterion is a triplet defined as (field, operand, value) where "field" is the name of some attribute of a security, a portfolio, or a user; "operand" is a symbol indicating what condition is being tested; and "value" is a fixed value that represents a requirement or threshold. The criterion (Expense Ratio, >-, 1%) represents the statement "The expense ratio of the security in question is at least 1%." This criterion might be used, for example, to indicate whether to prune funds in a certain investment category. One of skill in the art would quickly recognize other examples of attributes of securities that could be used within the context of the present invention. In the course of determining an optimized portfolio, tax optimization may be considered by the present system. For example, to alleviate the tax burden on a portfolio the system may be set to balance the short term losses against the short term gains within a portfolio. The present system may achieve this through recommending the sale of an investment over time, recommending a sale or purchase of an investment product at a later time. These rules are fully configurable to satisfy the investment style and preferences of the investment firm.

Theifirst step of the portfolio construction analysis algorithm of the present system is to prune the user's existing portfolio of securities deemed unacceptable. The ■ Pruning Rules table allows this process to be customized by an investment firm. Depending on the needs of the operator of the present system, the system can be configured to prune securities very aggressively or hardly at all. The present system preferably employs pruning rules to identify changes in the mutual fund portion of an investment portfolio. In the presently-preferred embodiments of the present invention, pruning rules are also employed to generate suggestions for investment in all categories of financial products. The pruning rules preferably establish a set of criteria by which investment products are evaluated.

Each investment category, user category, and portfolio partition (t.e. taxable or tax- advantaged) may have its own set of pruning rules. If a security meets any of the criteria specified in the appropriate rule set, that security is pruned. Note that the presence of the portfolio partition as a column in the table could cause certain securities to be pruned in taxable accounts, but not pruned in tax-advantaged accounts.

The above rule says to prune any U.S. Large Growth fund from a mid net worth user's taxable account that has an expense ratio of 1.5% or higher. Note that if a rule is not applicable to a particular security (for example, EXPENSE_RATIO does not apply to individual stocks), the rule is ignored for that security.

The recommendation rules of the present invention are refinements of the pruning rules in that if a fund would have been pruned, it will not be recommended for further investment. The criteria used in the context of pruning rules would mirror those used in the recommendation rules.

Following a pruning analysis of a portfolio, augmentation rules are then applied. A portfolio is augmented with a new investment product in any recommended category for every account where that investment category is not well represented by the holdings currently in the account. An investment category is considered well-represented in an account if the account's holdings in that investment category are sufficiently diversified to produce (if funded within the appropriate proportions) a portfolio with high enough expected return and low enough risk (variance) for that investment category. Investment product category averages are preferably used to determine the minimum return and maximum variance cutoffs.

An Augmenting Rules Table is used to define the conditions under which the present system will attempt to recommend new investment products (e.g. mutual funds, individual stocks, annuities, etc.) to the user. For each account in the user's portfolio, the present system will attempt to ensure that every recommended investment category is well represented in that account. For example, given an account A and an investment category C, the present system will preferably identify a set of securities (S) in investment category C, all of which are available to be included in the target portfolio. The set of securities S must satisfy the criteria established in the augmenting rules table for that particular category.

Note that by ensuring the possibility of investing in every investment category in every account, the present system in no way ensures that the target portfolio will actually have assets invested in such a way. The present system attempts to provide the user with as much freedom as possible to choose an optimized portfolio that spans all of the user's accounts.

The augmenting criteria R that the security set S must meet are to be interpreted in the following way. There must be a way to allocate funds across the securities of S such that R applies to the resulting portfolio. If there is no such solution, the present system will attempt to augment A with new funds (adding these funds to S) until the criteria can be met. The following table provides an example of augmenting rules.

Typically, the augmenting rules will be used to ensure that it is possible to achieve a particular level of return at a particular level of risk for each investment category.

Once the present system determines that a portfolio should be augmented, specific securities are selected from recommended securities. Each recommended security has a Recommendation Allocation table. The Recommendation Allocation table contains a recommended investment category breakdown for each user category and for each asset class. The breakdown is represented as a minimum percentage that should be allocated to that investment category in the target portfolio.

The above example provides that the target portfolio should have at least 18% of its stock investments in U.S. Large Blend stocks and/or funds.

Within a single asset class (e.g., stocks) it is required that the minimum percentages add up to no more than 100%. However, it is permissible for these percentages to add up to strictly less than 100%. If the system is in fact configured such that the percentages add up to less than 100%, the optimizer will have some flexibility in determining the precise breakdown of investment categories within each asset class.

Investment categories that are included in this table (with a minimum percentage greater than zero) are considered to be recommended investment categories. The remaining investment categories are considered to be non-recommended investment categories.

Once the system has determined that augmentation of a portfolio is necessary for one or more investment categories by using the rules exemplified above, it selects new investment products to recommend from the Recommended Securities Table. This table lists all of the funds that the system may recommend for a particular investor. The list is preferably ordered so that the system will choose the most highly recommended investment product in the appropriate investment category that is available in any particular account. The order of investment products to be recommended may be influenced directly by the investment firm in order to recommend particular investment products preferentially.

The Recommended Securities Table may be rebuilt as part of the periodic securities data processing script. The script may contain a fixed set of funds to put in the recommended securities table or it may contain an algorithm that analyzes the funds' fundamental and/or returns data in order to determine which funds to recommend. A fund's presence or rank in the recommend list may depend on the type of account (taxable or tax-advantaged) and/or the user category.

When presenting a set of recommendations to a user, the present system also preferably provides for the ability to inform the user of the reasons for the recommendations. In this context, the present invention is capable of generating text messages that accompany the investment recommendations. These messages preferably include an explanation of the reasoning behind the recommendations. For example, the present system may inform a user that he or she should reduce their investment in a particular mutual fund because the expense ratio cost is too high. Any particular explanation warranted by the particular implementation of the rules of the present system may be presented.

Although the invention has been described in terms of particular embodiments in an application, one of ordinary skill in the art, in light of the teachings herein, can generate additional embodiments and modifications without departing from the spirit of, or exceeding the scope of, the claimed invention. Accordingly, it is understood that the descriptions herein are proffered only to facilitate comprehension of the invention and should not be construed to limit the scope thereof.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A method for using a computer to form a recommendation for a portfolio, comprising the steps of:

receiving a financial goal;

receiving a description of an investment strategy;

receiving a description of an investment portfolio;

analyzing the investment portfolio to identify attributes of individual investment products;

using the financial goal and the investment strategy to generate an asset allocation for an investor;

producing a set of constraints based on the asset allocation, the attributes, and the investment strategy;

forming the recommendation for the portfolio by using the set of constraints; and

outputting the recommendation.

2. The method of Claim 1, wherein said analyzing step includes analyzing a return on investment of said individual investment product.

3. The method of Claim 1, wherein said analyzing step includes analyzing a capitalization of said individual investment product.

4. The method of Claim 1, further comprising the step of generating pruning rules and augmenting rules, after said producing step.

5. The method of Claim 4, further comprising the step of applying said pruning rules and said augmenting rules to said investment portfolio, after said generating step.

6. The method of Claim 1, wherein said step of forming the recommendation for the portfolio is carried out by using the investment attributes.

7. The method of Claim 1, wherein the step of analyzing the investment account to identify account attributes includes analyzing the investment account to identify a set of investment constraints.

8. The method of Claim 1, wherein said attributes of individual investment products include expense ratio costs, expected return on investment, and capitalization.

9. A method of using a computer to generate recommendations for an investment portfolio, comprising the steps of:

receiving information about an investor;

analyzing said information about an investor to generate an asset allocation;

establishing a set of rules, wherein said set of rules is used to analyze an investment portfolio;

receiving information about an investment portfolio, said investment portfolio comprising investment products;

analyzing said information about said investment portfolio to obtain attributes of said investment products;

evaluating said investment portfolio using said attributes of said investment products;

establishing a set of rules; and

generating a set of recommendations for investment by employing said set of rules using said asset allocation as a constraint.

10. The method of Claim 9, wherein said set of information about an investment portfolio includes a list of investment products in which said investor has invested.

11. The method of Claim 9, wherein said set of information about said investor includes a list of investment products in which said investor is presently invested.

12. The method of Claim 9, wherein said attributes of said investment products include an expense ratio cost, a capitalization, and an expected return on investment.

13. The method of Claim 9, wherein said investment product comprises stocks, bonds, cash, and annuities.

14. The method of Claim 9, wherein said set of rules includes pruning rules and augmenting rules.

15. The method of Claim 9, wherein said asset allocation comprises a percentage distribution of assets in stocks, bonds, and cash.

16. The method of Claim 9, wherein said set of rules includes tax optimization rules.

17. A method of using a computer to generate recommendations for an investment portfolio, comprising the steps of:

establishing a first database, wherein said first database comprises attributes of investment products;

establishing a second database, wherein said second database comprises attributes of a portfolio, wherein said portfolio comprises a set of investment products in which an investor has invested;

generating a list of investment products to be pruned by evaluating said second database using a set of pruning rules;

generating a list of investment products to be augmented by evaluating said second database using a set of augmenting rules;

generating a list of investment products to be invested in by evaluating said first database using a set of recommending rules, wherein said list of investment products to be pruned, said list of investment products to be augmented, and said list of investment products to be invested in comprise a set of recommendations;

delivering said set of recommendations to a user; and

generating a message that explains a reason for said recommendations.

18. The method of Claim 17, further comprising obtaining a set of information about said investor, before the first establishing step.

19. The method of Claim 18, further comprising generating an asset allocation based on said set of information about said investor, after the obtaining step.

20. The method of Claim 17, further comprising receiving an asset allocation, before the first establishing step.