Method and apparatus for producing time variant asset allocation

[0001] The present invention relates to methods and apparatus for producing a time variant asset allocation among a plurality of investments and, more particularly, to producing an allocation equation that may be utilized to predict a substantially optimal allocation of assets among the investments at one or more points in time.

[0002] It is desirable to determine an optimal allocation of assets among a plurality of investments (i.e., an investment portfolio). For example, an investor may wish to distribute his assets among investments A, B, and C. The investor's return on the portfolio will depend on the respective market values of investments A, B, and C, as well as the distribution of his assets by percentage among these investments (i.e., the allocation of his assets). It is self evident that the investor would like to maximize his return on the investment portfolio by selecting an advantageous allocation of assets among the investments.

[0003] In keeping with the desire to maximize the return on an investment portfolio, those skilled in the art have sought to develop procedures for determining an advantageous allocation of assets among a plurality of investments. For example, the so-called Markowitz model was developed in the early 1950s to compute a desirable allocation of assets among a plurality of investments based on historical relationships among the investments. More particularly, the Markowitz model is frequently implemented by requiring that the average returns of the respective investments and the standard deviations of those returns are computed for a particular historical period. A correlation matrix is then determined, which defines the extent to which the investments are linked (i.e., correlated) in terms of their market values over the historical period. The Markowitz model then uses a quadratic programming routine to compute an asset allocation among the investments that minimizes the square of the standard deviation of the returns of the investment portfolio. Inputs to the quadratic programming routine include a desired average return (for the investment portfolio set by the investor), the average returns for each investment, the standard deviations of these returns, and the correlation matrix. The resultant asset allocation is fixed as a function of time.

[0004] Unfortunately, the asset allocation obtained via the Markowitz model has significant drawbacks. For example, the Markowitz asset allocation does not provide an asset allocation that is time variant. Consequently, the investor must either use a fixed asset allocation and hope for the best over time, or recompute the average rates of return for each investment, the standard deviation of these returns, and the correlation matrix to determine a new asset allocation for a new time period. The new asset allocation, however, would be heavily skewed by the historical average of the returns of each investment and, therefore, would not provide satisfactory asset allocations, particularly for short term distributions (e.g., monthly, weekly, daily, etc.).

[0005] Further disadvantages of the Markowitz model include that it does not permit other market factors to affect the asset allocation and, therefore, the computed asset allocation cannot be influenced by, for example, leading market indicators. By way of example, many investments may be affected by inflation rates and, thus, it would be beneficial to adjust asset allocations based on them. Since the Markowitz model relies heavily on the historic performance of the portfolio investments (e.g., the average return), the Markowitz model has no mechanism for directly adjusting the asset allocation based on changes in current inflation rates. This could result in highly undesirable asset allocations when there is a significant disparity between average and current inflation rates over a relevant historical period. For example, a particular inflation rate may have averaged ten percent during the relevant historical period, but the current inflation rate may be three percent. The Markowitz model, however, would at best yield an asset allocation corresponding to the ten percent level.

[0006] Another model was developed by Konno and Yamazaki in the early 1990s to compute asset allocations among a plurality of investments. In their process, the historical monthly returns for each investment of the portfolio are used in a linear programming routine to minimize a sum of differences between the rates of return of the investments and an minimum desired rate of return. Like the Markowitz model, the Konno and Yamazaki model yields an asset allocation that is time invariant. Thus, the asset allocation computed by the Konno and Yamazaki model represents an average allocation for use in long term investing. The Konno and Yamazaki model is not equipped to provide an investor with the information needed to make short term asset allocation changes, such as monthly, weekly, daily, etc.

[0007] Accordingly, there is need in the art for new methods and apparatus for determining time variant asset allocations among a plurality of investments based, among other things, on market factors such that the investor can quickly respond to changing market conditions.

[0008] In accordance with one or more aspects of the present invention a method or apparatus is operable to facilitate the steps of establishing a plurality of investments in a portfolio among which an allocation of assets is to be made; establishing one or more factors to be associated with the plurality of investments, at least one of the factors having values that are time variant; and computing the allocation of assets among the investments as one or more functions of the one or more factors such that a time variant allocation of the assets among the investments is obtained.

[0009] Other aspects, features, advantages, etc. of the present invention will be apparent to one skilled in the art in view of the description herein taken in conjunction with the accompanying drawings.

[0010] For the purpose of illustrating various aspects of the invention, there are shown in the drawings forms that are presently preferred, it being understood, however, that the invention is not limited to the precise arrangements and instrumentalities shown.

[0011] FIG. 1 is a block diagram illustrating one example of a system suitable for computing an allocation of assets among a plurality of investments of a portfolio in accordance with one or more aspects of the present invention;

[0012] FIG. 2. is a partial flow diagram illustrating process steps and/or functions that may be utilized to compute the asset allocation in accordance with one or more aspects of the invention;

[0013] FIG. 3 is an example of at least part of a setup screen that may be utilized in accordance with one or more aspects the invention to receive data and/or instructions from an investor in order to produce an asset allocation;

[0014] FIG. 4 is an example of a report screen containing historical data for the investments of a portfolio suitable for use in accordance with one or more aspects of the invention;

[0015] FIG. 5 is an example of a report screen containing historical data for one or more factors that may be used to compute the asset allocation in accordance with one or more aspects of the invention;

[0016] FIG. 6 is a further partial flow diagram illustrating further process steps and/or functions that may be utilized to compute the asset allocation in accordance with one or more aspects of the invention;

[0017] FIG. 7 is an example of at least part of a further screen that may be utilized in accordance with one or more aspects of the invention to receive data and/or instructions from the investor in order to produce an asset allocation;

[0018] FIG. 8 is an example of a suitable data file containing information that may be used to compute the asset allocation in accordance with one or more aspects of the invention;

[0019] FIG. 9 is a further partial flow diagram illustrating further process steps and/or functions that may be utilized to compute the asset allocation in accordance with one or more aspects of the present invention;

[0020] FIG. 10 is an example of a report screen containing information that may be utilized to define one or more allocation equations for the respective investments of the portfolio in accordance with one or more aspects of the invention;

[0021] FIG. 11 is an example of a report screen containing data showing historical and/or predicted asset allocations for the investments of the portfolio in accordance with one or more aspects of the present invention;

[0022] FIG. 12 is an example of a report screen containing graphical information concerning the historical return on the investments of the portfolio in accordance with one or more aspects of the present invention; and

[0023] FIG. 13 is an example of a report screen containing data concerning statistical performance information for the investments of the portfolio in accordance with one or more aspects of the present invention.

[0024] Referring now to the drawings, wherein like numerals indicate like elements, there is shown in FIG. 1 a block diagram illustrating an example of a system 100 for computing an allocation of assets among a plurality of investments of a portfolio in accordance with one or more aspects of the present invention. The system 100 preferably includes a data processing unit 102 and a memory 104 operatively coupled by way of a data and/or instruction bus 106. The data processing unit 102 may be implemented utilizing any of the known hardware, such as a digital and/or analog microprocessor, a computer (such as a portable, a stationary and/or a distributed computing system), or any of the other known and/or hereinafter developed processing units. The memory 104 preferably includes an investment history database 108 and a factor history database 110. The memory 104 may be implemented by way of separate hardware or may be disposed within the data processing unit 102. Any of the known hardware and/or software for implementing the databases 106 and/or 110 may be employed, such as the commercially available CodeBase 6 database engine that may be utilized with a computer.

[0025] Data are preferably input to, and output from, the data processing unit 102 by way of an input/output device 112 that is operatively coupled to a display/data input device 120 via a data and/or instruction link 114. By way of example, the display/data input device 120 may include a display screen, such as any of the commercially available CRTs, LCDs, etc. The display/data input device 120 may also include any of the commercially available input devices, such as a keyboard, a mouse, a voice recognition system, etc.

[0026] A user of the system 100, such as an investor, preferably utilizes the display/data input device 120 to provide information to the data processing unit 102 over the data and/or instruction link 114 to facilitate the computation of one or more asset allocations (and/or the computation of one or more asset allocation equations) in accordance with the invention. The data processing unit 102 preferably provides information concerning the portfolio, such as the allocation of assets, to the investor by way of the display/data input device 120. A storage medium 122, such as a magnetic storage medium, an optical storage medium, etc. may also be utilized to provide information to, and receive information from, the data processing unit 102, e.g., by way of an appropriate storage medium reading device (not shown).

[0027] It is noted that the term investor herein is used broadly to include one or more individual investors, one or more institutional investors, their agents and their representatives, etc.

[0028] It is noted that the functional blocks illustrated in FIG. 1 may be partitioned as shown or may be partitioned in any other way, such as in an integral fashion. By way of example, the system 100 may be implemented utilizing a portable, stationary, or distributed computer operating under one or more suitable computer programs. Further, one or more of the functional blocks of the system 100 may be remotely located from the others, such as in a distributed (e.g., networked) system. For example, the display/data input device 120 may be remotely located from the other functional blocks of the system 100, where the link 114 may be an electronic communication network, such as the Internet.

[0029] Irrespective of how the system 100 is implemented and/or partitioned, it preferably carries out a method (e.g., a process) for computing an allocation of assets in accordance with one or more aspects of the invention such that a time variant allocation of the assets among the investments is obtained. To this end, reference is now made to FIG. 2, which is a partial flow diagram illustrating at least some steps and/or functions of this method. At action 200, the investor preferably selects a plurality of investments for the portfolio and makes this information available to the data processing unit 102. By way of example, and with reference to FIG. 3, the investor may provide information to the data processing unit 102 by way of a setup screen 300A that may be presented to the investor on the display/data input device 120. Any of the known processing techniques for producing the setup screen 300A may be employed without departing from the spirit and scope of the invention.

[0030] In the illustrated example, the setup screen 300A includes a plurality of fields, among them an investment input field 302, through which the investor may select the plurality of investments for the portfolio. In a first area 302A of the investment input field 302, a source of information from which the investments are selected is identified; in this case, the source is the user (or investor). The selected investments of the portfolio are preferably listed in a second area 302B of the investment input field 302. In this example, the investor has selected a stock index, a government bond, and the 90 day T-Bill. Preferably, the source of investment information may be changed by way of a drop-down menu box 302C and the selection of a particular investment may be made via a drop-down menu box 302D (together with an appropriate add, insert, or delete command, e.g., made using point and click techniques on activation areas 302E).

[0031] Referring again to FIG. 2, at action 202 a selection is preferably made as to a date range over which to analyze the historical performance of the investments of the portfolio. Preferably, this selection is made by the investor utilizing a setup screen, which may be the setup screen 300A of FIG. 3 or may be another setup screen. In this embodiment of the invention, the selection of the date range is preferably made by the investor using the setup screen 300B of FIG. 7. Further details as to how the investor uses the setup screen 300B of FIG. 7 (including how to select the date range) will be discussed later in this description.

[0032] At action 204 one or more factors are preferably established that are influential in predicting respective rates of return of the investments of the portfolio. As used herein, the term factor relates to any market factor, such as any macroeconomic factor (e.g., interest rate(s), inflation rate(s), GNP, unemployment rate(s), trade deficit, etc.), commodity prices, stock prices, bond prices, etc. Preferably the investor may select the factors and input them into the data processing unit 102 by way of the setup screen 300A (FIG. 3). More particularly, the setup screen 300A preferably includes a factor input field 304 including a first area 304A and a second area 304B that are respectively used to identify a source of the factor information and to specify the selected factors themselves. In this example, the following factors have been established: the S&P composite common stock dividend yield, the 10-year treasury rate, and the 1-year treasury rate.

[0033] As with the investment input field 302, the factor input field 304 preferably includes a drop-down menu box 304C for identifying and/or altering the source of the factor information, a drop-down menu box 304D for identifying and/or altering the selection of a particular factor, and add, insert, and delete selection areas 304E for facilitating the selection of the factors (e.g., using point and click techniques). The factor input field 304 also preferably includes a third area 304F in which to display the available data date ranges for each factor.

[0034] Any of the known techniques may be employed to determine whether a given factor should be used. By way of example, a determination as to the correlation (and lags) between a potential factor and one or more of the investments may be made (action 204A, FIG. 2). This determination may be made utilizing any of the known regression algorithms, such as those that employ auto correlation techniques, for quantifying relationships between respective time series. At action 204B, a determination is made as to whether the correlation obtained in action 204A is sufficiently high to suggest that the potential factor should be used (action 204C) or that the correlation is not sufficiently high (return to action 204A or end).

[0035] At action 206, the investment history and factor history over the date range of interest is preferably retrieved. With reference to FIG. 1, this historical data is preferably read from the investment history-database 108 and the factor history database 110. In accordance with the invention, the investment history data is preferably accessible and viewable by the investor via the display/data input device 120. By way of example, FIG. 4 illustrates a report screen 400 containing a tabulation of the respective returns of the investments of the portfolio on an historic basis (e.g., the monthly returns in 1972, 1973, 1974, etc.). Any of the known (or hereafter developed) techniques for generating the report screen 400 (and/or for generating further report screens discussed below) may be employed, such as using the commercially available Rogue Wave, Stingray Studio software program that is operable to run on known computers. Similarly, the investor is preferably permitted to access and view the factor history by way of the display/data input device 120, for example, by way of a report screen 402 (FIG. 5) containing a tabulation of the historical data concerning the factors. Preferably, the investor may readily select the reports 400, 402 by activating respective areas 400A and 402A (e.g., using point and click techniques).

[0036] With reference to the further flow diagram of FIG. 6, a variety of parameters utilized in computing the asset allocation(s) (and/or the asset allocation equation(s)) in accordance with the invention are preferably established. These include establishing a desired average portfolio return (action 208), establishing a threshold for a minimum rate of return for the portfolio (action 210), establishing minimum and maximum asset allocations for each of the investments of the portfolio (action 212), establishing transaction costs for each investment (action 213), establishing minimum and maximum allowable leverage ratios (action 214), establishing an objective function (action 216), and establishing constraint equations for determining the allocation of assets (action 218). Preferably, these parameters are established through user input (e.g., via the display/data input device 120, FIG. 1) or by way of an automated process executed by the data processing unit 102.

[0037] More particularly, and with reference to FIG. 7, the setup screen 300B (which may be separate from or may be a further portion of setup screen 300A) may be utilized to establish the parameters listed above. For example, setup screen 300B preferably includes an input area 306 into which the investor may enter a desired portfolio return, for example, a target monthly return of 1.00. The setup screen 300B also preferably includes a threshold input field 308 in which the investor may specify the threshold for minimum rate of return of the portfolio. It is noted that this threshold may be a constant, which is preferably specified by selecting area 308A and inputting a constant in area 308B, such as zero. Alternatively, the threshold may be specified by way of a time series by selecting area 308C and inputting a series of threshold values in area 308D, one value for each time period of interest.

[0038] As discussed above, the investor preferably selects and establishes the date range over which to analyze the historical performance of all of the investments of the portfolio (action 202 of FIG. 2) using the setup screen 300B. This is preferably done by inputting the date range into field 314, specifically area 314A. The area 314 includes an area 314B (which will be discussed later in this description) and an area 314C. Area 314C may be used by the investor to set the historical date range to the widest possible range based on the available date range of data (which is found at area 302F of FIG. 3).

[0039] The minimum and maximum asset allocations are preferably established by way of setup screen 300A (FIG. 3) via investor input at area 302G (e.g., a minimum allocation of zero percent (0%) and a maximum allocation of one-hundred percent (100%) for each investment).

[0040] The transaction costs (representing a percentage of each change in allocation) are preferably established by way of setup screen 300A via investor input at area 302H. A separate transaction cost can be established for each investment for transactions that add to allocations (Buy Cost) or decrease allocations (Sell Cost). These transaction costs represent such items as commissions, exchange fees, administrative fees, execution slippage and any other expense charged as a percentage of the value of an investment allocation change. If no fee is to be charged for a particular investment, the Buy Cost and Sell Cost can be established as zero. The transaction costs for all investments for all changes in allocation within one time period decrease the rate of return for the portfolio within that time period.

[0041] Establishing the minimum and maximum allowable leverage (action 214, FIG. 6) is preferably made by way of a leverage input field 310 of the setup screen 300B (FIG. 7). More particularly, the investor is preferably permitted to select whether leveraged investments are permitted by way of area 310A and, if so, entering a minimum leverage ratio and a maximum leverage ratio at areas 310B and 310C, respectively. When leveraged investment is used, the investor preferably enters a borrowing constant or a borrowing time series by selecting one of areas 310D and 310E, respectively, and entering the constant or series at area 310F or 310G.

[0042] The objective function (action 216, FIG. 6) is preferably established by way of investor input at field 312. Those skilled in the art will appreciate that the objective function may be a mathematical expression that represents a main goal for the portfolio, which goal is to be achieved through allocating the investor's assets among the investments of the portfolio.

[0043] In this example, the investor may select an objective function of minimizing deviations (by selecting area 312A) or of maximizing return by selecting area 312B. As used herein, minimizing deviations means substantially minimizing a sum of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio over time periods of interest. The objective function of maximizing return as used herein preferably means substantially maximizing an average of the rates of return of the portfolio over a plurality of time periods of interest. As will be discussed more fully hereinbelow, these are but two examples of suitable objective functions that may be used in accordance with one or more aspects of the invention. It is understood that other objective functions are contemplated without departing from the spirit and scope of the invention. For example, the objective function may be taken from the following group:

[0044] (i) substantially minimizing a sum of the differences between a rate of return of the portfolio and a minimum rate of return threshold for the portfolio over a plurality of time periods;

[0045] (ii) substantially minimizing a sum of squares of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio over a plurality of time periods;

[0046] (iii) substantially minimizing a variance of the rates of return of the portfolio over a plurality of time periods;

[0047] (iv) substantially maximizing a Sharpe ratio of the rates of return of the portfolio over a plurality of time periods; and

[0048] (v) substantially maximizing an average of the rates of return of the portfolio over a plurality of time periods.

[0049] Further details concerning these objective functions will now be provided. Turning to the first listed objective function (i), it is preferred that the sum of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio over a plurality of time periods, may be expressed as: $\left(1/M\right)\underset{t=1}{\overset{M}{}}\left(R_t-T_t\right),$

[0050] where M is a maximum number of time periods over which the sum of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio may be taken, R_t is the rate of return of the portfolio at time period t, and T_t is the minimum rate of return threshold for the portfolio at time t.

[0051] Preferably, the sum is computed only when (R_tT_t) is negative since, in this example, the goal for the portfolio (i.e., the objective) is to substantially minimize the sum of differences between the rate of return of the portfolio and the minimum rate of return threshold (i.e., when the rate of the return of the portfolio falls below the threshold). This may be achieved by permitting the difference of the rate of return of the portfolio and the minimum rate of return threshold for the portfolio at time period t, to be expressed as:

[0052] where Z_t is the difference of R_t and T_t at time period t when such difference is negative and zero otherwise, and V_t is the difference of R_t and T_t at time period t when such difference is positive and zero otherwise.

[0053] Turning to the second listed objective function (ii), it is preferred that the sum of squares of the rates of return of the portfolio, over a plurality of time periods, that fall below the minimum rate of return threshold for the portfolio, may be expressed as: $\left(1/M\right)\underset{t=1}{\overset{M}{}}{\left(R_t-T_t\right)}^2$

[0054] In this case, the sum is also preferably computed only when (R_tT_t) is negative.

[0055] Turning to the third listed objective function (iii), it is preferred that the variance of the rates of return of the portfolio over a plurality of time periods, may be expressed as: $\left(1/M-1\right))\underset{t=1}{\overset{M}{}}{\left(R_t-R_{\mathrm{avg}}\right)}^2,$

[0056] where M is a maximum number of time periods over which the sum of squares of rates of return of the portfolio may be taken, and R_avg is the average of the rates of return of the portfolio over the M time periods.

[0057] When the objective function involves the Sharpe ratio, the fourth listed objective function (iv), such ratio of the rates of return of the portfolio over a plurality of time periods is preferably expressible as: ${\left(R_{\mathrm{avg}}-\mathrm{RF}\right)/\left(\left(1/M-1\right)\right)\underset{t=1}{\overset{M}{}}{\left(R_t-R_{\mathrm{avg}}\right)}^2)}^{1/2},$

[0058] where RF is a substantially risk free interest rate available to an investor associated with the portfolio.

[0059] When the objective function involves substantially maximizing the average of the rates of return of the portfolio over a plurality of time periods, the fifth listed objective function (v), such computation is preferably expressible as: $R_{\mathrm{avg}}=\left(1/M\right)\underset{t=1}{\overset{M}{}}\left(\left(\underset{j=1}{\overset{N}{}}{\mathrm{AA}}_{\mathrm{jt}}{y}_{\mathrm{jt}}\right)+\left(1-{\mathrm{LEV}}_t\right)I_t\right),$

[0060] where N is a maximum number of the plurality of investments, AA_jt is an asset allocation given to a jth one of the investments during time period t, y_jt is a rate of return of a jth one of the investments at time period t, LEV_t is a leverage ratio employed during time period t, and I_t is an interest rate during time period t on money borrowed to leverage the portfolio.

[0061] When transaction costs are to be included in computing asset allocations, the objective function for R_avg is preferably modified to include positive or negative changes in the asset allocation for each investment between a previous time period and a current time period and a transaction cost for each investment associated with such positive or negative change in asset allocation. Preferably, the average return in this scenario is expressable as: $R_{\mathrm{avg}}=\left(1/M\right)\underset{t=1}{\overset{M}{}}\left(\left(\underset{j=1}{\overset{N}{}}{\mathrm{AA}}_{\mathrm{jt}}{y}_{\mathrm{jt}}\right)+\left(1-{\mathrm{LEV}}_t\right)I_t-\underset{j=1}{\overset{N}{}}{\mathrm{DP}}_{\mathrm{jt}}{\mathrm{CP}}_j-\underset{j=1}{\overset{N}{}}{\mathrm{DN}}_{\mathrm{jt}}{\mathrm{CN}}_j\right),$

[0062] where DP_jt is the absolute value of the change in allocation for investment j between time period t1 and time period t whenever such change in allocation is positive and zero otherwise, DN_jt is the absolute value of the change in allocation for investment j between time period t1 and time period t whenever such change in allocation is negative and zero otherwise, CP_j is the transaction cost for each unit of investment j for each transaction involving a positive change in the allocation to investment j, CN_j is the transaction cost for each unit of investment j for each transaction involving a negative change in the allocation to investment j, and the transaction costs for time period t1 are defined to be zero.

[0063] Preferably the constraint equations (action 218, FIG. 6) are established automatically by the data processing unit 102 in response to at least some of the previously entered parameters, such as the desired portfolio return (action 208), the threshold for minimum rate of return of the portfolio (action 210), the minimum and maximum asset allocations to each investment (action 212), the minimum and maximum allowable leverage (action 214), the selected objective function (action 216).

[0064] It is noted that the constraint equations are generally useful in ensuring that the asset allocation solution results in an achievable distribution (e.g., no negative allocations, etc.). Examples of useful constraint equations are preferably taken from the following group:

[0065] (i) that a leverage ratio employed during a given time period is substantially equal to a function of the values of the one or more factors during that time period;

[0066] (ii) that a rate of return of the portfolio in a given time period is substantially equal to a function of at least one of the asset allocations for each investment in that time period, rates of return of the investments in that time period, a leverage ratio employed during that time period, an interest rate during that time period on any money borrowed to leverage the portfolio, and transaction costs associated with the investments;

[0067] (iii) that a sum of the asset allocations for the investments is substantially equal to unity during a given time period;

[0068] (iv) that the sum of the asset allocations for the investments is substantially equal to the leverage ratio during the given time period;

[0069] (v) that an average rate of return of the portfolio over a plurality of time periods is substantially equal to an average of the rates of return of the portfolio in each of the plurality of time periods;

[0070] (vi) that the average rate of return of the portfolio over a plurality of time periods is substantially equal to a target average rate of return for the portfolio;

[0071] (vii) that the asset allocation for a given one of the investments is less than or substantially equal to a maximum permissible allocation for the investment as a proportion of all allocated assets;

[0072] (viii) that the asset allocation for a given one of the investments is greater than or substantially equal to a minimum permissible allocation for the investment as a proportion of all allocated assets;

[0073] (ix) that the leverage ratio employed during a given time period is less than or substantially equal to a maximum permissible leverage;

[0074] (x) that the leverage ratio employed during a given time period is greater than or substantially equal to a minimum permissible leverage;

[0075] (xi) that a sum of differences between a rate of return of the portfolio and a minimum rate of return threshold for the portfolio over a plurality of time periods, is less than or substantially equal to a maximum permissible average deviation below the minimum rate of return threshold for the portfolio;

[0076] (xii) that a sum of squares of differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio over a plurality of time periods, is less than or substantially equal to a maximum permissible average square deviation below the minimum rate of return threshold for the portfolio;

[0077] (xiii) that a variance of the rates of return of the portfolio over a plurality of time periods is less than or substantially equal to a maximum permissible variance; and

[0078] (xiv) that a Sharpe ratio of the rates of return of the portfolio over a plurality of time periods is greater than or substantially equal to a minimum permissible Sharpe ratio.

[0079] Further details regarding these constraint equations will now be given. Preferably, the constraint equation listed above at (i), i.e., that the leverage ratio employed during a given time period is substantially equal to a function of the values of the one or more factors during that time period, may be expressed as: ${\mathrm{LEV}}_t=C\underset{k=1}{\overset{K}{}}{F_{\mathrm{kt}}}^{}\left(P_k\right),$

[0080] where LEV_t is the leverage ratio employed during time period t, C is a constant associated with the leverage, and P_k represents a power for a kth one of the factors.

[0081] Most preferably, this constraint equation may be preferably expressed as: ${\mathrm{LEV}}_t=C+\underset{k=1}{\overset{K}{}}{D}_k{F}_{\mathrm{kt}},$

[0082] where LEV_t is the leverage ratio employed during time period t, C is a constant, and D_k is a coefficient associated with a kth one of the factors.

[0083] Preferably, the constraint equation listed above at (ii), i.e., that the rate of return of the portfolio in a given time period is substantially equal to a function of the asset allocations for each investment in that time period, rates of return of the investments in that time period, a leverage ratio employed during that time period, and an interest rate during that time period on any money borrowed to leverage the portfolio, may be expressed as: $R_t=\left(\underset{j=1}{\overset{N}{}}{\mathrm{AA}}_{\mathrm{jt}}{y}_{\mathrm{jt}}\right)+\left(1-{\mathrm{LEV}}_t\right)I_t,$

[0084] where AA_jt is the asset allocation for a jth one of the investments during time period t, N is a maximum number of the plurality of investments, y_jt is a rate of return of a jth one of the investments at time period t, LEV_t is a leverage ratio employed during time period t, and I_t is an interest rate during time period t to be paid on money borrowed to leverage the portfolio.

[0085] When transaction costs are to be considered in computing the asset allocations, the rate of return, R_t, is preferably modified to include positive or negative changes in the asset allocation for each investment between a previous time period and a current time period and a transaction cost for each investment associated with such positive or negative change in asset allocation. Preferably, the rate of return in this scenario may be expressed as: $R_t=\left(\underset{j=1}{\overset{N}{}}{\mathrm{AA}}_{\mathrm{jt}}{y}_{\mathrm{jt}}\right)+\left(1-{\mathrm{LEV}}_t\right)I_t-\underset{j=1}{\overset{N}{}}{\mathrm{DP}}_{\mathrm{jt}}{\mathrm{CP}}_j\underset{j=1}{\overset{N}{}}{\mathrm{DN}}_{\mathrm{jt}}{\mathrm{CN}}_j,$

[0086] where DP_jt is the absolute value of the change in allocation for investment j between time period t1 and time period t whenever such change in allocation is positive and zero otherwise, DN_jt is the absolute value of the change in allocation for investment j between time period t1 and time period t whenever such change in allocation is negative and zero otherwise, CP_j is the transaction cost for each unit of investment j for each transaction involving a positive change in the allocation to investment j, CN_j is the transaction cost for each unit of investment j for each transaction involving a negative change in the allocation to investment j, and the transaction costs for time period t1 are defined to be zero.

[0087] Preferably, the constraint equation listed above at (iii), i.e., that the sum of the asset allocations for the investments is substantially equal to unity during the given time period, may be expressed as: $\underset{j=1}{\overset{N}{}}{\mathrm{AA}}_{\mathrm{jt}}=1,$

[0088] where AA_jt is the asset allocation for a jth one of the investments during time period t, and N is a maximum number of the plurality of investments.

[0089] Preferably, the constraint equation listed above at (iv), i.e., that the sum of the asset allocations for the investments is substantially equal to the leverage ratio during the given time period, may be expressed as: $\underset{j=1}{\overset{N}{}}{\mathrm{AA}}_{\mathrm{jt}}={\mathrm{LEV}}_t,$

[0090] where AA_jt is the asset allocation for a jth one of the investments during time period t, N is a maximum number of the plurality of investments, and LEV_t is a leverage ratio employed during time period t.

[0091] Preferably, the constraint equation listed above at (v), i.e., that the average rate of return of the portfolio over a plurality of time periods is substantially equal to an average of the rates of return of the portfolio in each of the plurality of time periods, may be expressed as: $R_{\mathrm{avg}}=\left(1/M\right)\underset{t=1}{\overset{M}{}}{R}_t,$

[0092] where R_avg is the average of the rates of return of the portfolio over a plurality of time periods, M is a maximum number of time periods over which the sum of rates of return of the portfolio may be taken, and R_t is the rate of return of the portfolio in time period t.

[0093] Preferably, the constraint equation listed above at (vi), i.e., that the average rate of return of the portfolio over a plurality of time periods is substantially equal to a target average rate of return for the portfolio, may be expressed as: $R_{\mathrm{avg}}=R_{\mathrm{target}}=\left(1/M\right)\underset{t=1}{\overset{M}{}}{R}_t,$

[0094] where R_avg is the average of the rates of return of the portfolio over a plurality of time periods, R_target is a desired target average rate of return for the portfolio, M is a maximum number of time periods over which the sum of rates of return of the portfolio may be taken, and R_t is the rate of return of the portfolio in time period t.

[0095] Preferably, the constraint equation listed above at (vii), i.e., that the asset allocation for a given one of the investments is less than or substantially equal to about a maximum permissible allocation for the investment, may be expressed as:

[0096] where amax_j is a maximum permissible allocation as a proportion of all allocated assets for a jth one of the investments, and LEV_t is a leverage ratio employed during time period t.

[0097] Preferably, the constraint equation listed above at (viii), i.e., that the asset allocation for a given one of the investments is greater than or substantially equal to about a minimum permissible allocation for the investment, may be expressed as:

[0098] where amin_j is a minimum permissible allocation as a proportion of all allocated assets for a jth one of the investments, and LEV_t is a leverage ratio employed during time period t.

[0099] Preferably, the constraint equation listed above at (xi), i.e., that the sum of the differences between the rate of return of the portfolio and a minimum rate of return threshold for the portfolio over a plurality of time periods, is less than or substantially equal to a maximum permissible average deviation below the minimum rate of return threshold for the portfolio, may be expressed as: $\left(1/M\right)\underset{t=1}{\overset{M}{}}\left|Z_t\right|{\mathrm{DEV}}_{\max },$

[0100] where M is a maximum number of time periods over which the sum of rates of return of the portfolio may be taken, Z_t is a difference of the rate of return of the portfolio and the minimum rate of return threshold for the portfolio at time period t, and DEV_max is the maximum permissible average deviation below the minimum rate of return threshold. In this case, the sum of Z_t is computed only for negative values of Z_t(R_tT_t), where R_t is the rate of return of the portfolio at time period t, and T_t is the minimum rate of return threshold for the portfolio.

[0101] Preferably, the constraint equation listed above at (xii), i.e., that the sum of squares of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio over a plurality of time periods, is less than or substantially equal to a maximum permissible average square deviation below the minimum rate of return threshold for the portfolio, may be expressed as: $\left(1/M\right)\underset{t=1}{\overset{M}{}}{Z}_t^2{\mathrm{DEV}}_{\max }^2,$

[0102] where M is a maximum number of time periods over which the sum of squares of rates of return of the portfolio may be taken, Z_t is a difference of the rate of return of the portfolio and the minimum rate of return threshold for the portfolio at time period t, and DEV²_max is the maximum permissible average square deviation below the minimum rate of return threshold. In this case, the sum of Z_t² is computed only for negative values of Z_t(R_tT_t), where R_t is the rate of return of the portfolio at time period t, and T_t is the minimum rate of return threshold for the portfolio.

[0103] Preferably, the constraint equation listed above at (xiii), i.e., that the variance of the rates of return of the portfolio over a plurality of time periods is less than or substantially equal to a maximum permissible variance, may be expressed as: $\left(1/\left(M-1\right)\right)\underset{i=1}{\overset{M}{}}{\left(R_t-R_{\mathrm{avg}}\right)}^2{\mathrm{VAR}}_{\max },$

[0104] where M is a maximum number of time periods over which the sum of rates of return of the portfolio may be taken, R_t is the rate of return of the portfolio in time period t, R_avg is the average of the rates of return of the portfolio over a plurality of time periods, and VAR_max is the maximum permissible variance.

[0105] Preferably, the constraint equation listed above at (xiv), i.e., that the Sharpe ratio of the rates of return of the portfolio over a plurality of time periods is greater than or substantially equal to a minimum permissible Sharpe ratio, may be expressed as: ${\left(R_{\mathrm{avg}}-\mathrm{RF}\right)/\left(\left(1/M-1\right)\right)\underset{t=1}{\overset{M}{}}{\left(R_t-R_{\mathrm{avg}}\right)}^2)}^{1/2}{\mathrm{SHARPE}}_{\min },$

[0106] where M is a maximum number of time periods over which the sum of rates of return of the portfolio may be taken, R_t is the rate of return of the portfolio in time period t, R_avg is the average of the rates of return of the portfolio over a plurality of time periods, RF is a substantially risk free interest rate available to an investor associated with the portfolio, and SHARPE_min is the minimum permissible Sharpe ratio.

[0107] Referring again to FIG. 6, at action 220, a data file is preferably passed to a programming routine, such as a linear or quadratic programming routine in order to facilitate the computation of the asset allocation. Preferably, this data is passed by way of an electronic file, such as the electronic file illustrated in FIG. 8, and includes one or more of the following data: the date range, the desired portfolio return, the threshold for minimum rate of return, the minimum and maximum asset allocations for the respective investments, the buy and sell transaction costs, the minimum and maximum allowable leverage, the list of economic factors, the history of the factors over the date range of interest, the list of investments of the portfolio, the investment history over the date range of interest, information concerning the objective function, and information concerning the constraint equations.

[0108] By way of example, a linear programming routine is preferably employed when the objective function involves substantially minimizing the sum of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio. Conversely, a quadratic programming routine is preferably employed when the objective function involves substantially minimizing the sum of squares of the differences between the rate of return of the portfolio and the rate of return threshold for the portfolio. Any of the known linear or quadratic program routines may be employed without departing from the spirit and scope of the invention. For example, the commercially available Lingo 7.0 linear/non-linear programming package may be employed, which runs on conventional computers. Preferably, the electronic data file (FIG. 8) is tailored to provide the requisite data to the linear or non-linear programming routine; in this example, the electronic data file is preferably tailored for the Lingo 7.0 programming software.

[0109] With reference to the further flow diagram of FIG. 9, at action 222, the linear or non-linear programming routine preferably is executed by the data processing unit 102 to compute values for use in an allocation equation, which may be used to compute a time variant allocation of assets among the investments as a function of the factors. Preferably, the one or more allocation equations for each of the investments may be expressed as:

[0110] where j is an index representing each of the plurality of investments, t is an index representing time periods, AA_jt is an asset allocation for a jth one of the investments in time period t, k is an index representing each of the one or more factors, and F_kt represents a value of a kth one of the factors at time period t.

[0111] More particularly, the one or more allocation equations for each of the investments may be expressed as: ${\mathrm{AA}}_{\mathrm{jt}}=A_j\underset{k=1}{\overset{K}{}}{F_{\mathrm{kt}}}^{^}\left(P_{\mathrm{kj}}\right),$

[0112] where A_j is a constant of proportionality associated with a jth one of the investments, K is a maximum number of the one or more factors and P_kj represents a power for a kth one of the factors and a jth one of the investments.

[0113] Most preferably, the one or more allocation equations for each of the investments may be expressed as: ${\mathrm{AA}}_{\mathrm{jt}}=A_j+\underset{k=1}{\overset{K}{}}{B}_{\mathrm{kj}}{F}_{\mathrm{kt}},$

[0114] where A_j is a constant associated with a jth one of the investments, K is a maximum number of the one or more factors and B_kj represents a coefficient for a kth one of the factors and a jth one of the investments.

[0115] In this case, the linear programming routine preferably computes the constant A_j and the coefficients B_kj for each of the one or more factors. As noted above, the one or more factors include historical values and, therefore, the constant and the coefficients of the above allocation equations are preferably computed as functions of the historical values of the factors. It is noted that the computation of the constant A_j and the coefficients B_kj are preferably determined in a way that tends towards optimizing the objective function (such as substantially minimizing the sum of the differences between the rate of return of the portfolio and the minimum rate of return threshold for the portfolio). More particularly, the constant A_j and coefficients B_kj are most preferably computed in a way that tends towards optimizing the objective function and satisfying the one or more constraint equations. It is most preferred that the linear programming routine (or the non-linear programming routine) is operable to compute the constant A_j and the coefficients B_kj such that positive and/or negative values are permitted.

[0116] When the investor has established that leveraged investments may be permitted, then it is preferred that the constraint equations include at least one of the leverage constraint equations discussed hereinabove. By way of example, the constraint equation listed above at (i) may be employed, i.e., that the leverage ratio employed during a given time period is substantially equal to a function of the values of the one or more factors during that time period. Recall that this constraint may be expressed as: ${\mathrm{LEV}}_t=C+\underset{k=1}{\overset{K}{}}{D}_k{F}_{\mathrm{kt}},$

[0117] where LEV_t is the leverage ratio employed during time period t, C is a constant, and D_k is a coefficient associated with a kth one of the factors.

[0118] When such a constraint equation is utilized, the programming routine preferably simultaneously computes the constant A_j, the coefficients B_kj, the constant C, and the coefficients D_k. Again, it is preferred that the programming routine is operable to compute the constant A_j, the coefficients B_kj, the constant C, and the coefficients D_k such that positive and negative values are permitted.

[0119] Referring again to FIG. 9, the constant A_j, the coefficients B_kj, the constant C, and the coefficients D_k are preferably stored (e.g., in memory 104, FIG. 1) for concurrent or later use (action 224). Preferably, the investor is permitted to view these values by way of the display/data input device 120. For example, and with reference to FIG. 10, the data processing unit 102 preferably provides a report screen 404 to the display/data input device 120 that contains a tabulation of constant A_j, the coefficients B_kj, the constant C, and the coefficients D_k as related to the investments of the portfolio and the factors. It is noted that the report screen 404 is preferably made available to the investor when area 404A is activated by the investor (e.g., via point and click techniques). More particularly, the report screen 404 preferably lists the factors in area 404B and lists the investments at area 404C. Leverage information is preferably found at area 404D. It is noted that in this example, the minimum and maximum leverage ratios were set to 1.00, i.e., no leveraged investment was contemplated. Accordingly, the coefficients D_k are all zero.

[0120] Turning to a first one of the investments of the portfolio, e.g., the S&P stock index, the value of the constant A₁ was computed to be 0.3341, the value of coefficient B_1,1 was computed to be 0.2275, the value of coefficient B_2,1 was computed to be 0.0206, and the value of coefficient B_3,1 was computed to be 0.1198. The asset allocation for this investment may be placed into the form ${\mathrm{AA}}_{\mathrm{jt}}=A_j+\underset{k=1}{\overset{K}{}}{B}_{\mathrm{kj}}{F}_{\mathrm{kt}},$

[0121] as follows:

[0122] Using the above form for A_jt, the asset allocations for the other investments, namely, the long term government bond and the 90 day T-Bill may be expressed respectively as follows:

[0123] Referring again to FIG. 9, at action 226, a date range of interest for forecasting a desired asset allocation among the investments is established. Preferably, this date range is obtained by way of investor input, e.g., by entering the date range into the setup screen 300B (FIG. 7) at area 314B. By way of example, the date range may be entered by year and month, although any other time increments may be employed without departing from the scope of the invention.

[0124] At action 228, the asset allocation among the investments of the portfolio are preferably computed for the forecasted date range using the allocation equations. More particularly, the values of the factors at the time of interest are entered into the asset allocation equations to compute the desired allocations. In this sense, the asset allocation is time variant because the allocation is a function of the time variant values of the factors.

[0125] It is noted that the way in which the values of the factors are utilized in the asset allocation may be adjusted. For example, the beginning and ending lags of the factor values may be specified by inputting them at area 304G of the setup screen 300A (FIG. 3). Further, a particular operator may be specified at area 304H. These operators may include (i) an average, (ii) a sum, (iii) a minimum, (iv) a maximum, (v) compound, and (vi) a product.

[0126] With reference to FIG. 11, the investor is preferably permitted to view the time variant asset allocation by way of a report screen 406 displayed on the display/data input device 120. The report screen 406 preferably includes the date/time period at area 406A, a list of the investments and their respective asset allocations at area 406B, the rate of return on the portfolio at area 406C, and further information at area 406D, such as the minimum rate of return threshold, the leverage ratio, and the cost of borrowing money for leveraged investing. As can be seen in area 406A, the time period of interest may include historical time periods as well as future (forecasted) time periods.

[0127] The investor is preferably permitted to obtain other reports, such as a portfolio growth report screen 408 shown in FIG. 12. By way of example, the growth report screen 408 is an historical (and/or forecasted) graph showing the past and/or future growth of the investment portfolio. Again, the investor is preferably permitted to view the growth report screen 408 on the display/data input device 120.

[0128] Preferably, the investor is also permitted to access to a statistical performance report screen 410. e.g., by way of the display/data input device 120 as shown in FIG. 13. The statistical performance report screen 410 preferably includes a number of statistical quantities for each of the investments as well as the portfolio as a whole. These statistical quantities preferably include a one month arithmetic average, a 12 month arithmetic average, a one month geometric average, a 12 month geometric average, a highest month, a 95% high, a 90% high, a 90% low, a 95% low, a lowest month, a one month mean absolute deviation, a one month standard deviation, a 12 month standard deviation, a maximum draw down, a risk free interest rate, the Sharpe ratio, etc.

[0129] In accordance with one or more further aspects of the present invention, the methods and/or functions discussed hereinabove for computing the asset allocation among the investments of the portfolio (and/or computing the asset allocation equations) may be achieved utilizing suitable hardware, such as that illustrated in FIG. 1. Preferably, such methods are achieved by employing a processor that is operable to execute instructions of one or more software programs. The one or more software programs preferably cause the processor (e.g., the data processing unit 102 of FIG. 1) to execute the actions and/or functions discussed hereinabove with respect to FIGS. 2-13. The one or more software programs are preferably operable to be stored on any of the known or hereinafter developed storage media, such as magnetic storage media, optical storage media, electronic storage media, floppy disks, optical disks, memory chips, etc. For example, the one or more software programs may be stored on storage medium 122 (FIG. 1). Advantageously, the one or more software programs may be easily transported and/or distributed to investors or other users by way of the storage media.

[0130] Advantageously, the methods and apparatus discussed hereinabove with respect to FIGS. 1-13 permit an investor to compute time varying asset allocations among the investments of the portfolio such that changes to the asset allocation may be readily computed and executed. This provides the investor with a significant advantage over prior art techniques in, for example, maximizing the return on the investments of the portfolio.

[0131] Although the invention herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. It is therefore to be understood that numerous modifications may be made to the illustrative embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims.