JP2005234614A - Asset allocation status display device, asset allocation status display method, program, and screen - Google Patents

Abstract

【Task】
To make it easy to understand differences or similarities between portfolios created due to risk quantification models and data characteristics used when creating portfolios.
[Solution]
The display portfolio specification section allows the user to specify the portfolio to be displayed by its identifier or attribute value, and the distance definition specification section allows the user to specify the method for calculating the distance between portfolios and the attribute value used for distance calculation. The distance relationship display section displays the portfolios with specified attribute values based on the definition of the distance between the specified portfolios so that each portfolio is displayed such that the distance relationship between each portfolio is maintained. It is a status display device.
[Selection] Figure 1

Description

  The present invention relates to asset management / management technology.

  As asset allocation methods, conventionally known methods are used to allocate assets that maximize returns under a certain risk in consideration of the risk-return trade-off, or assets that minimize risk to obtain a certain return. It has been.

  In addition, the risk and return of each investment target brand and product are estimated from past price trend data, and each investment is made to minimize (or maximize) the risk (or return) of the entire asset composed of them. Decide the allocation rate of the target brands and products. These asset allocations are called portfolios, and determining a portfolio as described above is called portfolio optimization.

  Japanese Laid-Open Patent Publication No. 2002-06357 discloses a portfolio creation system and method as a conventional technique using comparison between portfolios or extraction of similarity between portfolios. In Japanese Patent Laid-Open No. 2002-06357, in relation to trading of stocks in a portfolio of a plurality of investors, the similarity between the brands of each investor's portfolio and the brands to be traded is extracted, and trading using the same This is technology related to portfolio creation support. However, it is not intended to support the portfolio creation of each investor, and only the type of issue is analyzed for the effects of various factors such as differences in risk measurement models considered in the portfolio creation process. There is nothing aimed at.

JP 2002-06357 A

  In the conventional portfolio optimization method, it is necessary to estimate the risk and return of each investment brand and product from each past price transition data. Risk is the possibility of not getting a return or incurring a loss due to a decrease in the price of each investment target. However, there are various methods for measuring risk. There are a method of expressing the magnitude of the return fluctuation in standard deviation, a method of measuring only the loss side of the return fluctuation distribution, and a method of paying attention to the expected value of the huge loss. Moreover, since the risk value by these is an estimated value based on the past price transition data of each investment object, it is influenced also by the quality of past data (data length, the presence or absence of a missing side), etc.

  When portfolio optimization is performed using the amount of risk measured by these various methods and data quality, the allocation ratio of each investment object in the resulting portfolio is different. Each portfolio reflects the characteristics of the risk quantification method used and the quality of the data, and the investment planner must determine the portfolio that seems to be the best from their own perspective by comparing each portfolio.

  However, it is possible to visually understand the characteristics of each portfolio's investment allocation ratio and to understand the relationships and differences between the portfolios, and what factors are responsible for these differences. It is not easy to grasp. It is a problem to realize an asset allocation display method or an asset allocation display apparatus that efficiently supports grasping and portfolio determination.

  An object of the present invention is to provide an asset allocation state display device, an asset allocation state display method, a program, and a screen capable of visually grasping the characteristics of each portfolio.

  One feature of the present invention is that the asset allocation display device includes an asset allocation data storage unit, a display portfolio designation input unit, a distance definition designation input unit, and a distance relationship display unit between portfolios.

  According to the present invention, the characteristics of each portfolio can be grasped visually.

  An example of the configuration of an asset allocation display device according to the present invention is illustrated in FIG. FIG. 2 illustrates a processing flow performed by the central processing unit 101 in FIG. 1 using each unit. Processing performed by each unit in FIG. 1 will be described according to the processing flow of FIG.

  First, data relating to a portfolio is stored in the asset allocation data storage unit 102 (process 201). FIG. 3 shows an example of storing asset allocation data. It is assumed that a plurality of portfolios to be stored have already been created by performing optimization calculation using a data set and one risk quantification model. Here, the stored data is a portfolio No. identifying each portfolio. 302, asset allocation 301 of each portfolio, risk quantification model and data characteristics 305 used at the time of creating each portfolio. The asset allocation ratio 301 is an investment allocation ratio 304 in each portfolio with respect to the brands 1 to N of each investment target. In addition, as the risk quantification model 307, here, standard deviation, C-VaR, and absolute deviation, which are three types of risk indicators used for risk measurement, are used as model names. In addition, as data characteristics, here, when raw data is used for past time-series data (data supplementation: none) and when processed data subjected to data supplement processing is used (data supplementation: yes) Yes.

  Here, the risk quantification model 307 and the data characteristics can be said to be attribute values of the asset allocation state.

  In addition, as other attribute values, for example, a return estimation model or an asset type may be arbitrarily set as a data item.

  After the asset allocation data is stored in the asset allocation data storage unit 102 (process 201), the user inputs information for designating a portfolio to be displayed as an analysis target from the display portfolio designation input unit 103 (process 202). . FIG. 4 shows an example of an interface screen that the display portfolio designation input unit 103 has. In FIG. 4, as the portfolio to be analyzed and displayed, the direct portfolio No. 401 may be designated by a check button 404 or the like, or all portfolios having stored attributes may be designated. For example, if the C-VaR model of the risk metric model item 402 is checked, all portfolios having the C-VaR model as an attribute in the asset allocation data shown in FIG. 3 are designated. The same applies to the data supplement item 403. FIG. 4 shows a state in which all risk measurement models and portfolios without data supplementation are designated.

  After the process 202, the user further inputs a distance index used for analysis display and a method for calculating the distance from the distance definition designation input unit 104 (process 203). FIG. 5 shows an example of an interface screen that the distance definition designation input unit 104 has. In FIG. 5, a region 501 corresponds to a distance index designation input, and a region 502 corresponds to a distance calculation method. Here, as the distance index, the asset allocation ratio of each portfolio to the investment target, the largest historical loss of the portfolio, and the sharp ratio (return / risk) representing the efficiency of the portfolio are prepared as the distance index. Two items are specified in the box (503).

  The distance index can be specified as a plurality or a single distance index.

  Further, as a distance calculation method, a square distance and a Mahalanobis distance are prepared, and the square distance is designated by a check box (504). Here, the Mahalanobis distance represents the distance between a group and a point (or a group and a group, or a point within the group), while the normal distance (Euclidean distance) is the distance between the points. It is. Measured by taking into account the variation (covariance) of points within a group, it becomes a measure of how close a point is to the reference group (whether included or not), and in cluster analysis to classify sample points, etc. Well used.

  After processing 202 and 203, based on the designated input information, the inter-portfolio distance relationship display unit 105 uses the designated distance index value for the designated portfolio to determine the distance between each portfolio. Then, it is calculated by the designated distance calculation method, and a display is made so as to save the magnitude relation between these distances (process 204). FIG. 6 shows an example of a display method of the distance relationship display unit 105 between portfolios. In FIG. 6, each portfolio is plotted in a planar shape with the maximum loss on the horizontal axis and the asset allocation on the vertical axis among the distance indices specified on the screen example of FIG. 5.

Each portfolio may be distinguished and plotted by changing the shape of plot points depending on the combination of designated attributes. In FIG. 6, according to the attribute designation of the risk metric model item 402 and the data supplement item 403 in FIG.
(603), the C-VaR model is plotted with Δ (604), and the absolute deviation model is plotted with ○ (605).

In FIG. 6, for example, the portfolio 604 (using the C-VaR model) and the portfolio 605 (using the absolute deviation model) have the same maximum loss value, but their asset allocation is considerably different. On the other hand, in the portfolio 604 (using the C-VaR model) and the portfolio 603 (using the standard deviation model), the asset allocation rate is similar, but the maximum loss value is quite different due to the subtle difference. .

  The axis of the plot plane in FIG. 6 may select any one of the specified distance indexes, and without specifying the axis, the distance relationship reflecting all index values can be displayed without the axis display. Plotting may be performed, and various display methods as described above may be switched on the display screen of FIG. Further, a method of displaying three-dimensionally by specifying three axes can be considered.

  After the processing 204, the distance relationship display unit 105 between the portfolios waits for a confirmation input from the user as to whether or not to classify the portfolio on the display plane based on the display method at that time. If not, the process proceeds to processing 207.

When performing classification, the portfolio classification unit 106 classifies the portfolio according to the display method employed by the distance relationship display unit 105 between the portfolios at that time, and displays the classification information on a display plane (process 207). ). FIG. 7 shows a display example of portfolio classification information. FIG. 7 shows an example in which the portfolio plotted by the display method shown in FIG. 6 is classified and the classification information is displayed on the same plot plane. In FIG. 7, as a result, each risk quantification model is classified, and the points in the classification group are connected and distinguished by an identification line. In the example of FIG. 7, when the standard deviation model (group A) and the absolute deviation model (group B) are used as risk quantification models, the relationship between the maximum loss and asset allocation is similar, and C−
It can be seen that the VaR model (group C) is quite different. The distance between these groups is displayed using a matrix table such as 708, for example.

  For the distance between groups, a distance calculation method used in a clustering method widely known as a group classification method can be applied. In addition, it is also possible to classify groups that are close to each other by the above clustering method without specifying a classification index such as a risk quantification model as described above, or to enclose all such groups by surrounding them with a circle. An arbitrary attribute for clustering may be specified, for example, the number of elements in the group, an average distance, or the like may be specified. In addition, the definition of the distance for clustering is specified in process 203. It may be equivalent to the definition of distance.

  After the process 206, the user can change the portfolio to be displayed or change the definition of the distance to repeatedly analyze the display from various viewpoints. In that case, the process returns to 202, and otherwise, the process ends (process 207).

  As described above, according to the present invention, the user switches the designation by the display portfolio designation unit and observes the planar plot, thereby changing the portfolio asset allocation ratio according to the risk calculation model and the data set. Changes due to reliability can be easily grasped. In addition, the user can easily grasp the degree of difference / similarity from the same viewpoint of the same asset allocation by switching the designation using the distance definition designation unit and observing the planar plot. Furthermore, the user can easily grasp the similarity between groups, risk calculation model, or similarity between data sets from various viewpoints by observing the grouping by the asset allocation classification unit in a planar manner. .

1 is a configuration example of an apparatus for implementing the present invention. An example of the processing flow for implementing this invention. An example of the contents and storage method of asset allocation data. An example of a designated input screen for a portfolio to be displayed. An example of a designation input screen for defining a distance. Display example of distance relationship between portfolios. Display example of portfolio classification information.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 ... Central processing unit, 102 ... Asset allocation data storage part, 103 ... Display portfolio designation input part, 104 ... Definition definition input part of distance, 105 ... Distance relation display part between portfolios, 106 ... Portfolio classification part.

Claims (11)

An asset allocation data storage unit that stores a plurality of asset allocation states;
A display portfolio designating unit that receives the asset allocation state selected by the user among the plurality of asset allocation states;
A distance definition designation input unit that accepts a user to select an index and a calculation method of the distance between the selected asset allocation states;
An asset allocation state display device comprising: a distance relationship display unit between portfolios that displays the selected asset allocation state based on the selected distance index and a calculation method. In claim 1,
An asset allocation state display device comprising a portfolio classification unit for grouping related portfolios and displaying the classification information. In claim 1,
The asset allocation state display device, wherein the asset allocation state is represented by an optimal asset allocation result and a risk metric model. In claim 1,
An asset allocation state display device characterized by accepting an asset allocation and risk measurement model name selected by a user in a display portfolio designation section. In claim 2,
An asset allocation state display device characterized in that the shape of a plot point is changed for each group. In claim 2,
An asset allocation state display device characterized in that a distance between the groups is displayed. In claim 1,
An asset allocation state display device characterized in that the display is made in a two-dimensional or three-dimensional manner based on the selected index and calculation method of the distance. In claim 1,
An asset distribution state display device characterized by switching and displaying various display methods designated by a user. Means for receiving the asset allocation state selected by the user among the plurality of asset allocation states stored in the asset allocation data storage unit;
Means for accepting a user to select a distance indicator and calculation method between the selected asset allocation states;
An asset allocation status display method, comprising: means for displaying the selected asset allocation status based on the selected index of the distance and a calculation method. On the computer,
Means for receiving the asset allocation state selected by the user among the plurality of asset allocation states stored in the asset allocation data storage unit;
Means for accepting a user to select a distance indicator and calculation method between the selected asset allocation states;
A program for realizing means for displaying the selected asset allocation state based on the selected index of the distance and a calculation method. A screen that displays an asset allocation state based on a predetermined distance index and a predetermined calculation method.

Images