KR20140005606A - Method and system for asset management using portfolio control - Google Patents

Abstract

The present invention relates to a method and system for asset management using portfolio control capable of controlling financial factors of a portfolio by using control engineering and allowing a user to seek a result exceeding a benchmark market index and to seek a positive profit rate regardless of the market situation. A system for asset management using portfolio control includes a control unit sensing an error signal corresponding to a difference between a target financial factor predetermined for a portfolio and a total estimation financial factor of the portfolio and outputting a control signal for enrollment ratio decision for each item to minimize the error signal based on an individual estimation financial factor and a change rate of an individual estimation financial factor to time for each item in the portfolio, an enrollment ratio adjustment unit changing the investment proportion of each item based on the control signal and adjusting an enrollment ratio of each item based on the changed investment proportion, and a financial factor estimation unit sensing the enrollment ratio of each item from the portfolio, which changes its characteristics by enrollment ratio adjustment of the enrollment ratio adjustment unit and market situation, estimating an individual estimation financial factor and a change rate of the individual estimation financial factor to time in the enrollment ratio of each item, and generating a total estimation financial factor of the portfolio based on the individual estimation financial factor.

Description

Asset management system and method through portfolio control {METHOD AND SYSTEM FOR ASSET MANAGEMENT USING PORTFOLIO CONTROL}

The present invention relates to an asset management system and method through portfolio control, and more particularly, to control the financial variables of the portfolio to the desired value by using control engineering to achieve performance exceeding the benchmark market index. It also relates to asset management systems and methods that can seek positive returns regardless of market conditions and through portfolio control.

In investment theory, a portfolio generally refers to the investment target when diversifying investments in two or more assets such as stocks, bonds, and real estate, without investing in a single asset. Refers to configuration. In other words, in financial markets, a portfolio is a form of investment in which an economic entity determines the share of assets in consideration of risks and returns in order to maximize the utility of the assets in their portfolio. In addition, in relation to securities investments, an economic agent can pre-distribute uncertain risks in the future by diversifying investments in securities with different characteristics in their portfolios, which can be viewed as a portfolio in the stock market.

An example of an investment method using securities portfolio optimization in a stock market or futures market is disclosed in Korean Patent Publication No. 10-2009-0123351 (2009.12.02). The portfolio optimization method of this publication predicts changes in the baseline index based on the baseline index and the total daily trading value to determine when to buy and sell, and to determine which industries are included in the portfolio at which time they are purchased. After selecting the stocks, we propose an investment method that uses a genetic algorithm to determine the weight of the stocks in the portfolio so that the market capitalization of the portfolio follows the reference index. In addition, in the related art of the above-mentioned publication, a tracking portfolio can be minimized with a small amount of calculation by constructing a portfolio that determines a buying point and a selling point according to the correlation between the daily closing price of the reference index and the difference in the trading volume of the reference index. It is said to be able to maximize the trading profit of the portfolio.

However, the above-described prior art makes it possible for an economic entity to configure the weight or market cap of each item in its portfolio to follow the market cap of the stock market so that the trading margin of the portfolio is hard to escape the category of trading margin according to the market cap of the market. There is a limit. In particular, the portfolio optimization system described above has a structure that seeks a positive rate of return when the market situation is rising, but can only follow the negative rate of return when the market situation is down, so that a professional investment analyst can assist in investment analysis. It can be used as a portfolio control system for asset management.

As such, most investment analysis techniques based on conventional portfolio selection (HMMarkowitz, 1952) attempt to effectively follow the benchmark according to the market situation by optimizing the weight of the stocks in the portfolio. It was not possible to consistently pursue positive returns.

The present invention has been derived in view of the above-described limitations of the prior art, and an object of the present invention is to provide a portfolio using a control engineering technique so that the financial variables of the portfolio do not follow a market reference index but reach a predetermined desired value. By controlling the incorporation rate of each stock, it provides asset management system and method through portfolio control that can pursue performance exceeding the index and always pursue positive returns regardless of market conditions.

In order to solve the above technical problem, the asset management system through portfolio control according to an aspect of the present invention, an error signal corresponding to the difference between the target financial variable and the total estimated financial variable of the portfolio preset for the portfolio for asset management A control unit which detects and outputs a control signal for determining an inclusion ratio for each item such that an error signal is minimized based on an individual estimated financial variable for each item in the portfolio and a change rate over time of the individual estimated financial variable; An inclusion ratio adjusting unit for converting the investment ratio of each item based on the control signal and adjusting the inclusion ratio of each item in the portfolio based on the converted investment ratio; And re-detecting the inclusion ratio of each item from the portfolio whose characteristics are changed by the adjustment of the inclusion rate adjustment section and information input from the market, and removing the noise of newly inputted information from the inclusion ratio of each item. It includes a financial variable estimator that estimates the rate of change of a financial variable and an individual estimated financial variable over time, and calculates the total estimated financial variable of the portfolio based on the individual estimated financial variable. Here, the output of the financial variable estimator is input to the controller. And estimating the rate of change of the individual estimated financial variable over time corresponds to estimating the first derivative over time of the individual estimated financial variable.

The financial variable is an index indicating the sensitivity of the portfolio to the market, and may be at least one of a beta coefficient, a correlation coefficient, a volatility, and a delta coefficient. Here, the beta coefficient is a financial variable that represents the sensitivity of the portfolio to the market changes, the correlation coefficient is a financial variable that measures the degree of relationship between the financial variable and the market index, the volatility is constant A financial variable that represents the extent to which a price of a term stock, bond, or commodity fluctuates. The delta coefficient is a financial variable that changes the price of an option against changes in underlying asset prices.

These financial variables may vary in the length of time needed to statistically calculate the financial variables. For example, in the case of the beta coefficient, the same item appears different depending on how much basic data, such as 5, 10, 20, 60, 120, 120, etc., to calculate the beta coefficient, in the present invention The control of financial variables includes all the calculation periods applicable to the market regardless of the calculation period described above, and the user may select and apply the financial variables of the calculation period according to the market situation and other factors.

In one embodiment, the inclusion ratio adjustment unit may perform the function for the following two purposes. The first objective is to adjust the inclusion ratio so that the portfolio's performance exceeds the benchmark (eg benchmark market index). In that case, the weight of each item is determined to establish a long position based on the target financial variable and the operation method through the control of the system, and the feedback rate is established so that a new long position is established every time the whole process is repeated by feedback. Can be determined. The second objective is to adjust the inclusion rate so that the portfolio's performance does not exceed the performance of the benchmark market index, but always achieves a positive rate of return, regardless of whether the overall market rises or falls. In that case, create an agenda index by converting 100% of the actual share of each item in the portfolio to the market, calculate the individual agenda share for each item in the agenda index, and then target financial as in the first objective. Subtracting the weight of the agenda for the same item in the previously created agenda index from the weight of the long position of each individual item created based on the variable and the operating method through the control of the system, the long position or short position according to the individual item This is determined, and a final Market Neutral Long / Short Position is established in which the net position representing the directionality of the long and short positions in the entire portfolio is zero. Here, if the above process is repeated through feedback as new information of the market is inputted to the portfolio, the inclusion ratio is determined so that a market neutral long short position is established every time the feedback is repeated, thereby increasing the overall market. And a positive rate of return can always be achieved regardless of the decline.

In one embodiment, the controller is implemented as a Proportional Integral Derivational (PID) controller. In addition, the controller may be implemented as any kind of controller having a function of minimizing an error between the actual financial variable or the total estimated financial variable and the target financial variable.

In one embodiment, the financial variable estimator is implemented with a Kalman filter. In addition, the financial variable estimator is a filter having a function of estimating a financial variable, removing noise of time series financial information (financial variable, market information, etc.) or estimating the rate of change of an individual estimated financial variable over time. Available. For example, there are various types of filters such as low-pass filter, high-pass filter, and band-pass filter, and other high-performance filters are continuously being developed. . Considering this filter development trend, the present invention uses the Kalman filter, but the present invention is not limited to such a configuration, and is at least one filter which performs a function corresponding to the Kalman filter in addition to the Kalman filter, and is marketed and individual items. You can use any filter with the ability to remove noise from the information and estimate the rate of change of the financial variable over time.

In one embodiment, the asset management system through portfolio control is to change the rate of change over time of the overall market index (KOSPI 200, etc.) to effectively calculate the rate of change of the time of each variable received from the financial variable estimator. Estimate the target financial variable to move faster than the market movement when the rate of change of the market index shows a positive value, or equal to the movement of the market when the rate of change of the market index shows a negative value over time. It further includes a target financial variable setting unit for specifying the target financial variable to move slower than the market movement.

In the asset management method through portfolio control according to another aspect of the present invention, in the asset management method for controlling the portfolio in a feedback control method through an asset management system having a control unit, an integrated ratio adjustment unit and a financial variable estimating unit, Detecting an error signal corresponding to a difference between a predetermined target financial variable and a total estimated financial variable of the portfolio for the portfolio for asset management (first step); Outputting a control signal for determining an inclusion ratio for each item such that an error signal is minimized based on the individual estimated financial variable for each item in the portfolio and the rate of change over time of the individual estimated financial variable (second step); Converting the investment ratio of each item based on the control signal (third step); Performing at least one of buy, sell, short sell, and buyback for each item to adjust the transfer ratio of each item in the portfolio based on the converted investment ratio (fourth step); Adjust the incorporation ratio of each item and re-detect the incorporation rate of each item from the portfolio whose characteristics are changed by the information input from the market, and at the time of individual estimated financial variable and individual estimated financial variable Generating a rate of change for the second step (step 5); And calculating a difference between the total estimated financial variable of the portfolio based on the individual estimated financial variables and the target financial variable of the preset total portfolio (sixth step). Here, the sixth step is implemented to start a new loop in combination with the first step through a feedback process.

In one embodiment, the asset management method through portfolio control is that the sensitivity of the change in the portfolio is greater than the movement of the market when the rate of change of the overall market index (eg, KOSPI 200) over time before the first step is positive. Set the target financial variable so that the sensitivity of the portfolio change is less than or equal to the movement of the market when the rate of change over time of the overall market index is negative. Include.

In an embodiment, the fourth step may include calculating an incorporation ratio of each item by a target financial variable and a control operation method when pursuing performance exceeding the benchmark market index. Further, in an embodiment, the fourth step may include: calculating an actual weight ratio of each asset class in the market cap in the market, when always seeking a positive rate of return regardless of the market situation; Calculating an agenda weight (or agenda index) for each asset class by converting the actual weight to 100%; And recalculate the inclusion ratio of each item in the portfolio so that the overall estimated financial variable of the portfolio can follow the target financial variable, or based on the difference between the recalculated inclusion ratio and the agenda weight ratio; Final finalization rate).

In one embodiment, the fifth step includes estimating the rate of change of the individual estimated financial variable and the individual estimated financial variable over time using a Kalman filter.

According to the present invention, by controlling the incorporation rate of each item belonging to the portfolio by using the control engineering technique so that the financial variables of the portfolio do not follow the market reference index, but the predetermined value, it is always positive regardless of the market situation. It is possible to provide an asset management system and method through portfolio control that can achieve a return of more than and a performance exceeding the reference index.

In addition, according to the asset management system and method through the portfolio control of the present invention, in any market worldwide, if there is sufficient liquidity and benchmark index, it always seeks a positive rate of return regardless of the investment method and the market situation that exceed the performance. It can provide a solution on how to invest.

1 is a schematic block diagram illustrating an asset management system through portfolio control according to an embodiment of the present invention.
FIG. 2 is a schematic flowchart of an asset management method through portfolio control employable in the asset management system of FIG. 1.
FIG. 3 is a graph for explaining an embodiment of a tuning process for an incorporation ratio controller of the asset management system of FIG. 1.
4 is a graph schematically showing the control performance of the asset management system of FIG.
5 and 6 are graphs illustrating the distribution of financial variables before and after noise removal by the financial variable estimating unit of the asset management system of FIG. 1.
7 is a graph illustrating the performance of the portfolio using the asset management system of FIG.
FIG. 8 is a view for explaining a process of generating an agenda index employable in the asset management system of FIG. 1.
FIG. 9 is a view for explaining a process of determining an inclusion ratio of each item in a portfolio in the asset management system of FIG. 1.
10 is a graph illustrating the return on investment using the asset management system of FIG.

Hereinafter, preferred embodiments of the asset management system and method through the portfolio control of the present invention will be described in detail with reference to the accompanying drawings.

The terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to the intention or convention of a user or an operator. Therefore, definitions of these terms should be made based on the contents throughout this specification.

In addition, the following examples are not intended to limit the scope of the present invention, but merely illustrative of the components set forth in the claims of the present invention, which are included in the technical spirit throughout the specification of the present invention and constitute the claims Embodiments that include a substitutable component as an equivalent in the element may be included in the scope of the present invention.

1 is a schematic block diagram illustrating an asset management system through portfolio control according to an embodiment of the present invention.

Referring to FIG. 1, the asset management system 100 through portfolio control according to the present embodiment includes a control unit 10, an integrated ratio adjusting unit 20, and a financial variable estimating unit 30.

는 포트폴리오의 목표 재무변수를 의미하고,

는 바로 직전에 추정된 포트폴리오의 재무변수 또는 전체 추정 재무변수를 의미하며,

는 목표 재무변수에서 추정된 포트폴리오의 재무변수를 차감한 오차, 오차 값 또는 오차 신호를 의미한다.In Figure 1

Means the target financial variable of the portfolio,

Means the financial variable of the immediately preceding portfolio or the total estimated financial variable,

Means the error, error value, or error signal by subtracting the estimated financial variable from the target financial variable.

The controller 10 detects an error signal corresponding to the difference between the target financial variable preset for the portfolio P for asset management and the total estimated financial variable of the portfolio, and the individual estimated financial variable for each item belonging to the portfolio and A control signal is determined to determine the inclusion ratio for each item so that the error signal is minimized based on the rate of change of the individual estimated financial variables over time. Such a control unit 10 can be implemented using various types of general purpose controllers used in control engineering. In one example, the controller 10 may be implemented as a Proportional Integral Derivative (PID) controller.

The target financial variable preset for the aforementioned portfolio is the financial variable of the portfolio set to a desired value by the system user (hereinafter, simply referred to as a user) for the portfolio. In addition, the overall estimated financial variable of the portfolio is the financial variable or portfolio of the current portfolio based on the target financial variable and the feedback signal (current inclusion ratio of each item in the portfolio) that changes according to the market situation in the asset management system 100. As the current financial variable of, we remove the noise from the actual financial variables of each stock observed in the market and correspond to the weighted average of the individual estimated financial variables for each stock.

Incorporation ratio adjustment unit 20 converts the control signal into the investment ratio of each item, and adjusts the inclusion ratio of each item of the portfolio based on the converted investment ratio.

The inclusion ratio adjustment unit 20 may be implemented as a means for outputting a signal corresponding to the new inclusion ratio of each item in order to adjust the investment ratio of the items belonging to the portfolio. In this case, the portfolio P performs at least one of buy, sell, short and repurchase for each item through a trading system for various buy, sell, short sell or red buy in the market.

The financial variable estimating unit 30 incorporates each stock from the portfolio (hereinafter referred to as an investment portfolio) whose characteristics are changed by the adjustment of the inclusion ratio in the inclusion ratio adjustment unit 20 and information input from the market (market situation, etc.). Is detected again and noise is removed from the inclusion ratio of each item to generate individual estimated financial variables. In addition, the financial variable estimating unit 30 calculates the rate of change of the individual estimated financial variables over time and the total estimated financial variables of the portfolio based on the individual estimated financial variables. The output of the financial variable estimating unit 30 is input to the control unit 10.

)과 각 종목의 개별 추정 재무변수의 시간에 대한 변화율(

)이 입력되면, 제어부(10)는 미리 설정된 제어 프로세스에 따라 오차 값이 최소화되도록 각 종목의 개별 추정 재무변수를 기준으로 포트폴리오의 각 종목에 대한 편입 비율을 조정하기 위한 제어 신호를 출력한다. 제어 신호가 입력되면, 편입비율 조정부(20)는 제어 신호에 기초하여 개별 추정 재무변수로부터 각 종목에 대한 투자 비중을 환산하고, 환산한 투자 비중에 기초하여 포트폴리오의 각 종목의 편입 비율을 산출한다. 산출된 각 종목에 대한 새로운 편입 비율(

)은 투자 포트폴리오(P)에 전달된다. 이렇게 산출된 편입비율(

)과 시장에서 새롭게 입력되는 정보(주식시장의 경우, 주가 정보)에 의해 포트폴리오의 특성이 변하기 때문에 위와 같은 과정을 계속 반복하는 형태로 제어 동작이 수행된다. 여기서, 포트폴리오의 특정 종목의 편입비율이 제어 전의 제1 편입비율(

)에서 제어 후의 제2 편입비율(

)로 조정되는 과정에서 포트폴리오의 손익이 발생하게 된다.In this embodiment, the previously calculated error value (

) And the rate of change over time of each estimated financial variable for each stock (

) Is input, the control unit 10 outputs a control signal for adjusting the inclusion ratio for each item in the portfolio based on the individual estimated financial variable of each item in order to minimize the error value according to a preset control process. When the control signal is input, the inclusion ratio adjusting unit 20 converts the investment ratio for each item from the individual estimated financial variables based on the control signal, and calculates the inclusion ratio of each item in the portfolio based on the converted investment ratio. . New transfer rates for each of the calculated stocks (

) Is transferred to the investment portfolio (P). The transfer ratio calculated in this way (

) And the newly entered information in the market (in the case of the stock market), the characteristics of the portfolio change, so the control operation is performed in the form of repeating the above process. Here, the transfer ratio of a specific item of the portfolio is the first transfer ratio before control (

), The second incorporated ratio after control (

The profit and loss of the portfolio is generated in the process of adjusting to).

The aforementioned financial variables represent the sensitivity of the portfolio to the market, and beta coefficients, correlation coefficients, volatility, and delta coefficients are available. Here, the beta coefficient means the sensitivity of the portfolio to the market changes. A beta coefficient of 1 means that the portfolio moves as much as the market moves. It means that the ship moves more. These financial variables may have different values, even if they are the same financial variable, depending on the period of time over which the statistical data are obtained. For example, the values of the 100-day beta coefficient using the 100-day time series data and the 50-day beta coefficient using the 50-day time series data are different. If you want to control the 100-day beta coefficient of the entire portfolio, the beta coefficients of individual stocks will also use the values calculated from the 100-day time series data.

Meanwhile, in the present embodiment, the asset management system 100 adjusts the target financial variable such that the sensitivity of the change in the portfolio is greater than the movement of the market when the entire market is expected to rise based on qualitative or quantitative data. For example, a target financial variable setting unit (not shown) may be further provided to adjust a target financial variable such that the sensitivity of the portfolio change is less than or equal to the movement of the market when the entire market is expected to be in a downward phase. . In this case, the target financial variable setting unit sets the target financial variable to a preset value based on a signal for estimating the rate of change of the entire market from the financial variable estimating unit 30 and various data for judging other market conditions. It can be implemented to change.

FIG. 2 is a schematic flowchart of an asset management method through portfolio control employable in the asset management system of FIG. 1. FIG. 3 is a graph for explaining an embodiment of a tuning process in the controller in the asset management method of FIG. 2.

The asset management method through portfolio control with reference to Figures 2 and 3 as follows.

When the target financial variable for the portfolio is determined, the controller 10 detects an error signal corresponding to the difference between the target financial variable and the overall estimated financial variable of the portfolio through a predetermined input unit (S21). Here, the asset management system measures the actual beta coefficient of the portfolio through sensors or filters at regular intervals such as every minute, every hour or every day. When a difference occurs between the total estimated beta coefficient and the target beta coefficient obtained from the actual beta coefficient, the asset management system calculates the error through the input unit and transmits the calculated error to the controller 10 to detect it.

When an error or an error signal is detected, the controller 10 controls the inclusion ratio of each item (stock, etc.) of the portfolio to control the actual beta coefficient of the portfolio to approach the target beta coefficient. That is, the controller 10 outputs a control signal for determining the inclusion ratio of each item of the portfolio so that the error signal is minimized based on the change rate of the individual estimated financial variable and the individual estimated financial variable over time (S22).

When the control signal is input, the inclusion ratio adjusting unit 20 converts the investment ratio of each item based on the control signal (S23). Then, the inclusion ratio adjusting unit 20 adjusts the inclusion ratio of each item based on the investment ratio of each item (S24). Indeed, the adjustment of the inclusion ratio of each stock is carried out by buying, buying or selling short or repurchasing stocks in the investment portfolio.

When the inclusion ratio of each item is detected from the investment portfolio by a predetermined sensor or the like, the financial variable estimating unit 30 calculates primitive financial variables of each item based on the detected market and the price change of each item (S25). . In addition, the financial variable estimating unit 30 generates a change rate with respect to time of the individual estimated financial variable and the individual estimated financial variable of each item by removing noise from the financial variable originally calculated by the existing statistical calculation method ( S26). Then, the financial variable estimating unit 30 calculates the total estimated financial variables of the portfolio based on the individual estimated financial variables and compares them with the target financial variables in the next step (S27). The total estimated financial variable of the financial variable estimating unit 30 and the change rate for the predetermined period of the individual estimated financial variable and the individual estimated financial variable are input again to the controller 10 as a new input.

Meanwhile, in the above-described step (S22), an embodiment of the control unit 10 for minimizing the error will be described as a control operation of the PID controller.

First, when the PID controller is expressed as a transfer function, it is expressed by Equation 1.

here,

Respectively.

In order to implement the output signal of the PID controller, the output signal of the controller in the time domain is represented by the equation (2).

here,

Respectively.

On the other hand, the portfolio of the stock market has a characteristic that the characteristics of the market continue to change. Therefore, when a portfolio consisting of dozens or hundreds of stocks is used as a control target, it is very difficult to mathematically formulate it. Therefore, tuning of the coefficients of the PID controller must be performed. In this embodiment, Ziegler- Initial values of the coefficients are set on the basis of the Niegler-Nichols tuning method. By adjusting the coefficients of the PID controller through this simple tuning process, the coefficients of the PID controller can be determined simply.

The Ziegler-Nichols tuning method uses a proportional controller at first to increase the proportional control gain value from 0 and obtains a critical gain in which the vibration appears in the output as shown in FIG. It is a method to determine the coefficient value of the controller by obtaining period. In the Ziegler-Nichols tuning method, y (t) denotes the output of the plant, i.e., Pcr denotes the period of the vibration output.

Here, if the portfolio is managed by stocks and the funds are managed, the plant is the portfolio, the input of the plant is the transfer ratio allocated to each stock, and the output of the plant is the price of the individual stock and the beta coefficient of the portfolio. On the other hand, the state variable of the plant can be determined by the beta coefficient and the change rate of the beta coefficient of individual stocks according to market changes.

In addition, the detailed coefficient determination method according to the type of PID controller is shown in Table 1.

here,

Respectively.

Using Table 1, C (s), the transfer function of the PID controller, is expressed as Equation 3 below.

here,

Respectively.

In general, the control performance tends to be better as the shape of the plant is not changed over time with the basic characteristics of the plant under control. Plants in industrial sites have distinct shapes, such as cylinders driven by liquid or gas pressures or motors driven by electrical forces, and their characteristics do not change significantly over time. However, it is true that portfolios of stocks are not physically shaped and their characteristics always change over time, so it is unlikely that control performance will last for a long time.

However, even if the characteristics of the portfolio continue to change, they will remain in control for a short period of time, allowing full control over these short periods. Therefore, in the asset management system according to the present embodiment, the financial variable of the portfolio is controlled to a desired value by using the rate of change of each estimated financial variable of each item over a short period in which the control performance can be sufficiently exerted. By repeatedly performing such a task every fixed time, the actual financial variables (actual beta coefficients, etc.) are kept close to the target financial variables (target beta coefficients, etc.). The short time or the predetermined time means a period in which the characteristics of the portfolio are maintained and may be from minutes to days or months or more depending on the market situation.

4 is a graph schematically showing the control performance of the asset management system of FIG.

The procedure for controlling financial variables in the asset management system of this embodiment is as follows. For convenience of explanation, the case where the financial variable is a beta coefficient is described.

First, set the reference beta coefficient of the portfolio according to the market situation, and if the observed actual beta coefficient is lower than the target beta coefficient, increase the transfer ratio of stocks with high beta coefficient using the control unit, and vice versa. For example, we adjust the portfolio's beta coefficient to be closer to the target beta coefficient by increasing the transfer ratio of stocks with low beta coefficients.

In this process, the control output of the controller that controls the inclusion weights between the items is determined by a control error that is a difference between the observed beta coefficient and the target beta coefficient. If the control error is large, the adjustment of the incorporation ratio is made drastically because the control output is large. On the contrary, if the control error is small, the incorporation ratio is made less. This feedback is repeated every control cycle, bringing the actual beta coefficient of the portfolio closer to the target beta coefficient. The control output is determined by applying proportional control gain, differential control gain and integral control gain to the control error, the derivative value of the control error and the integral value of the control error, respectively.

In the present embodiment, the control ability of the beta coefficient depends on whether the actual beta coefficient is maintained at a level close to 1 when the target beta coefficient of the portfolio is set to one. If such control exists, a positive correlation between the beta coefficient and the share price return exists in a rising market, so the target beta coefficient is set to a level higher than 1, and in a down market, the beta coefficient and the share price return are negative. Since there is a negative correlation, the target beta coefficient can be set below 1 to seek abnormal returns that exceed the returns of the portfolio invested in terms of market cap.

The market situation can be predicted through various data, and qualitative prediction through research papers provided by securities companies can be made. For example, it can be quantitatively determined using the rate of change of the market index filtered through the Kalman filter. When determining the market situation using the Kalman filter, if the rate of change of the filtered market index shows a positive value, it is determined that the upward market will be maintained, so that the target beta coefficient is set higher than 1, and the filtered market If the rate of change of the index shows a negative value, it is determined that the bear market is maintained, and the target beta coefficient can be set to a level lower than 1.

For reference, the asset management system and method through the portfolio control of the present invention introduces the concept of market index to find out how excellent the present investment method and to define the range for the stock pool. For example, if the stock market in Korea, such as "KOSPI 200" may be the target. In other words, the meaning of the market index can be expressed as "indexed to express the characteristics of a certain market". The stock market index can be used as an indicator to gauge whether the stock market is going up or down as a whole. Do. The KOSPI 200 for the Korean stock market, the Nikkei 225 index for the Japanese stock market, the Dow Jones Index or the S & P 500 index for the US stock market can be used. In addition, the stock market in most countries has their own market indices, and in fact the index itself is used for trading in the market.

Also, market indices are not just in the stock market. There are bond indexes that show the characteristics of bonds well, and precious metals index, agricultural product index, general metal index, energy index, etc., which show the trend of various commodities such as gold, silver, copper, nickel, wheat, soybean, corn, oil, natural gas, etc. There are various indices traded in the market.

As such, the asset management method and system through portfolio control according to the present invention can be applied not only to the market index of the stock market but also to all market indices traded in the global market. In other words, in the present invention, the market index can be comprehensively applied to the market index using not only the stock market but all assets having investment value traded in the market.

In addition, the market index has the individual stocks that make up the market. In the case of KOSPI 200, this index is calculated using information on 200 stocks such as Samsung Electronics and POSCO. In the case of the KOSPI 200, the weight of each stock is determined based on the market capitalization of the liquidity of the stocks of each stock. In addition, representative stock market indexes such as Japan, the United States, and Europe, or bond indexes, agricultural product indexes, and energy indexes all have unique methods in which individual stocks that constitute the index and the weight of the individual stocks are determined. For example, in determining the weight of individual stocks, a methodology that is uniquely determined for each market index among various methods such as not only the market cap method, but also the weight based on the change in price and the same weighting method. This exists. In the asset management method and system through portfolio control according to the present invention, any one of the above-described methods may be applied to the method for determining the weight of individual items of the market index in addition to the method described herein.

In addition, regarding the method of setting the default pool of the portfolio, for example, in the case of KOSPI 200, 200 stocks are already named, the user of the asset management system of the present invention may select all 200 stocks as the default pool. Depending on the importance of the stocks, the default pool can be defined with only 150, 100, or 50 stocks within 200, except for the stocks that are not very important. When a pool of basic stocks is determined, weights of individual stocks are determined using the stock management system using the stocks in the pool. For example, in the KOSPI 200 index, the weight of Samsung Electronics is 20% in the current market situation, but the weight of Samsung Electronics may be 15% using the system or method of the present invention. In this way, the weightings of POSCO and Hyundai Motors are calculated separately according to the beta coefficient of the entire portfolio to be followed, unlike the weightings originally determined by KOSPI 200. As a result, the weight of individual stocks of the entire pool is determined by the method or system of the present invention according to the user's intention, and the performance of the pools, or portfolios, of the individual stocks thus determined is better than that of the KOSPI 200 index. Can be done.

As such, the asset management system and method through the portfolio control according to the present invention is better than the method of determining the individual weight, that is, the weight for the individual items, that the market index has through the control engineering operation method. More stable and better performance than yield. According to the present invention, it is possible to provide a solution for an investment method that can consistently yield better mid- and long-term results than the KOSPI 200 results.

The following describes the noise reduction and financial variable estimation process using the Kalman filter as an embodiment of the above-described financial variable estimator. 5 and 6 are graphs illustrating the distribution of short-term financial variables before and after noise reduction by the financial variable estimator of the asset management system of FIG. 1. In this embodiment, the case where the beta coefficient is used as the financial variable will be described.

The Kalman filter model is commonly used to estimate the state variables in a system by measuring the outputs that contain errors in the control system. The Kalman filter model is used to estimate the dynamic system state-space. It can be defined as an algorithm that continuously updates the linear projection of the system according to the sequential flow of time.

The Kalman filter model is constructed as follows. First, a system model and a measurement model may be represented by state functions such as Equation 4, Equation 5, and Equation 6.

here,

Respectively.

Here, w and v have no correlation with each other, and there is no correlation with x (0), and R is assumed to be greater than zero.

For a theoretical framework of the Kalman filter model described above, see Frank, L. Lewis (Optimal Estimation with an Introduction to Stochastic Control Theory, John Wiley & Sons, 1986), James, D. Hamilton (Time Series Analysis, Princeton University Press, 1994) and Andrew, C. Harvey (Forecasting, Structural Time Series Models and the Kalman Filter, Cambridge University Press, 2001).

Initializing the covariance and the estimated value of the state variable is shown in Equation 7.

here,

Respectively.

In the present embodiment, the covariance of the state variable and the time update of the estimated value due to the effect of system dynamics and the effect of measurement are determined as shown in Equations 8 and 9 do.

here,

Respectively.

In Equations 8 and 9, T indicates a transpose matrix, and 1 indicates an inverse matrix. In addition, the middle dot (·) indicates d / dt, which is a change rate with respect to time, and the circumflex accent (^) indicates an estimated value, not an observed value.

The Kalman gain is determined by the above equations as shown in equation (10).

here,

Respectively.

If the stock price formed in the stock market is regarded as the output from one system, it will actually contain various noises or errors. Thus, the Kalman filter can be used to remove this noise, while still estimating the state variables of the system over time. In the capital asset pricing model, the returns of individual stocks are represented by Equation 11, and the beta coefficient of individual stocks is determined by Equation 12.

In Equation 11 and Equation 12 above,

Respectively.

Meanwhile, if the state variable (x) in Equation 5 is defined as in Equation 13 and the state space function for the state variable (x) is set to be as in Equation 14, the state space function is represented by Equation 15 and Equation 14. 16, 17 and 18 can be configured.

In Equations 15 to 18,

Respectively.

Finally, the Kalman filter model is represented using the above equations. Equations 18 and 19 are estimated, and the estimation of the beta coefficient is updated by Equation 20, Equation 21 and Equation 22.

In the above equations 19 to 23,

Respectively.

As described above, the filtered beta coefficient and the rate of change of the beta coefficient can be obtained for all individual stocks that make up the portfolio. Can be represented as:

here,

Respectively.

In order to determine the gain of the Kalman filter of Equation 10, the variance of the noise must be determined. Since the mathematical modeling of the beta coefficient and the rate of change of the beta coefficient is very clear, there are almost no errors related to the system model. Can be seen (Q = 0). Therefore, the beta coefficient measured in the market does not accurately reflect the intrinsic beta coefficient of the stock at this point, and only the variance of noise related to the measurement needs to be estimated.

The variance of the measured noise can be estimated by the following method using the maximum likelihood estimation method. If the distribution of the measured beta coefficient follows a normal distribution, the probability density function may be expressed as Equation 26.

The likelihood function is the same as Equation 27, and the gain of the Kalman filter can be obtained by obtaining the variance of system noise maximizing the logarithm of the likelihood function (28).

here,

Respectively.

On the other hand, since the distribution of stock price or beta coefficient is not normally distributed in the real market, and most of the portfolio characteristics are not linear, the filter function that wants Kalman gain obtained by the maximum likelihood estimation method is performed. If it is set to a value for the purpose, the designer's intention may not be reflected. Accordingly, in another embodiment of the present invention, a procedure of first obtaining an initial value of a gain of a Kalman filter by using a maximum likelihood estimation method and then changing the gain little by little in the period of configuring the asset management system, and confirming the filtering effect according to the simulation. Repeat to determine the Kalman gain.

The filtering scale coefficient may be expressed as a ratio of the variability of the filtered signal to the variability of the unfiltered signal. Therefore, using Equation 29 may reflect the designer's intention to maintain the desired filtering performance at a desired level.

here,

Respectively.

As described above, when the beta coefficient is obtained by using a statistical method, as shown in FIG. 5, the beta coefficient is changed from about -5 to about 8 so severely that it is beyond the range applicable to reality. By applying the Kalman filter as shown in Fig. 6, the distribution of beta coefficients of individual stocks in the portfolio varies from about -1.0 to about 3 so that they are within the appropriate range available. Here, the beta coefficient is a short-term beta coefficient of 5 days to 10 days, which corresponds to an example.

In addition, in order to apply the beta coefficient to control engineering, it is necessary to capture the differential signal of the beta coefficient time, which also includes a lot of noise. Therefore, in the present embodiment, the Kalman filter is applied to the financial variable estimator to capture and use the first derivative signal, that is, the velocity signal of the beta coefficient, with respect to the time applicable to reality.

7 is a graph illustrating the performance of the portfolio using the asset management system of FIG.

This embodiment corresponds to the application of an asset management system through portfolio control to a mutual fund based on a long position. Long Position is a way of building a portfolio by buying stocks. It is an investment method that always pursues better performance than the market performance (cumulative return according to the comparative example). It is possible to pursue earnings that exceed benchmark returns (eg, the return of the Kospi200).

For example, we first set up a benchmark index (e.g., KOSPI200), select a stock pool based on stocks that contribute to the index among stocks in the stock market, and then, based on the stocks in the stock market, By setting the value of the target beta coefficient to be greater than 1 when the target beta coefficient is lower than 1 and by setting the value of the target beta coefficient to less than 1 when the market is in a downturn, the weight of each individual stock is always better than the market.

In addition, as shown in FIG. 7, the cumulative yields of the method according to the present embodiment and the method according to the comparative example (market capitalization ratio) increase as a whole during the investment period of a predetermined period (T1 to T13), and at the end of the investment period. The cumulative abnormal return according to the present embodiment is positively abnormal and increases after passing between T1 and T3 (about 1 month), and then decreases after about T9 and T11 (about 9 months). Overall, the cumulative abnormal return was maintained at about 12% / year (1% / month).

As described above, according to the present embodiment, asset management (investment, etc.) may be performed so as to follow the shape of the market as it is, but yield a better yield than the market.

On the other hand, unlike mutual fund investment, there is also Market Neutral Hedge Fund investment method that always seeks a certain amount of return (absolute return) regardless of the market situation. Hereinafter, a case where a market neutral hedge fund is implemented by applying the asset management system and method of the present invention will be described with reference to FIGS. 8 and 9. 8 is a view for explaining the calculation principle of the weight of individual items of the agenda index of the market. 9 is a view for explaining the principle of determining the ratio of inclusion of each item in the portfolio in the asset management system of FIG.

Market Neutral is one of the hedge fund's key strategies, not to exceed any benchmark index, but to always seek a certain amount of return, whether the market rises or falls. In order to create market neutrality in the portfolio, we do not always build Long Only positions like mutual funds, but rather long positions (buy) and short positions (short selling), and the net positions are almost zero. Always seek positive returns by getting closer.

For example, as shown in FIG. 8, the most influential items in the stock index are selected up to a cumulative weight of 95% of the index. For example, in the case of KOSPI 200, even if only the top 120 stocks are selected, it is close to 95% to 98% of the overall index (see 101). Then, if the portfolio of the selected stocks is converted to 100% in proportion to the index weight in the market, the weight of the individual stocks can be found to be almost the same as the individual weights of the original index (see 102). If the index obtained is made into an agenda index, the agenda index becomes very similar to the original index of the market, and the beta coefficient of the agenda index is almost one.

Next, as shown in FIG. 9, after setting a target beta coefficient for the portfolio including 120 items of the aforementioned agenda index 102, the weight of individual items of the portfolio calculated through the asset management system of the present invention ( Subtracting the weight of individual events (see 120) from the agenda index (see 110) can yield the final long short position (see 130) with a certain systematic risk (beta factor). For example, in the asset management system of the present embodiment, a portfolio having a target beta factor of 1.2 and subtracting the agenda index can create a market neutral long short position that pursues absolute profits with very small systematic risk of 0.2.

The Market Neutral Long / Short strategy is a trading technique that buys two assets and generates a short sale. This long-shot strategy is one of the hedge fund's representative strategies. Herein, hedge funds refer to specialized private equity funds. Private equity funds are funds that operate from a small number of investors, while internationally, private equity funds are mostly hedge funds.

On the other hand, if the long position and the short position are the same size, additional funding is not required if the collateral is secured, so the principal's principal is included in short-term liquidity stocks such as government bonds and Tongan bonds. The entire principal investment portion can be constructed to generate interest income. According to the long-shot market neutral hedge fund of this embodiment, it is possible to expect additional profits by subtracting expenses such as loan fees from interest from the principal invested. As such, investment or asset management according to the present embodiment may be implemented as a market-neutral long short position through an asset management method through bond purchase and portfolio control.

10 is a graph illustrating the return on investment using the asset management system of FIG.

FIG. 10 illustrates 120 events in a manner described above with reference to FIGS. 8 and 9, from a specific time point P1 (eg, March 20, 2006) to a specific time point P6 (eg, December 20, 2010). Is the result of the simulation. The securities transaction tax was taken into account during the simulation, and the loan cost and transaction fees paid to the prime broker were assumed to be treated as interest income (3-5%) of the principal invested bonds.

As shown in FIG. 10, the investment portfolio (the present embodiment) using the asset management system of the present invention shows a steadily upward return rate regardless of the fluctuation of the KOSPI (Comparative Example) for almost 5 years. The annual volatility during this period is 7.4% and the annual rate of return is 9.96%. Considering the return on volatility, it can be seen that the asset management system of the present invention works well for a long time.

As such, the success or failure of the hedge fund investment strategy is to achieve a stable return, and it is difficult to achieve long-term stable results while controlling volatility in the short term without systematic control logic. This was achieved through. In particular, according to the present embodiment, steady investments are being made not only for the Korean market but also for the global market with liquidity, thereby providing steady trust to domestic institutions and high-value asset owners, thereby expanding the global position beyond the leading position of the hedge fund market. We can provide asset management systems and methods that can be developed into hedge funds.

On the other hand, in the above-described embodiment has been described centering on the stock market index, the present invention is not limited to such a configuration. In other words, the asset management method and system through the portfolio control of the present invention can always pursue a positive rate of return regardless of the investment method and market that performs well in excess of liquidity and benchmark index in any market in the world. Can provide a solution to the investment method.

In addition, the asset management system and method through the portfolio control of the present invention, if only liquidity in the market is ensured, it is possible to adjust the key variables of the system to the relevant market based on the basic technical idea of the present invention to the global market simply. Applicable According to this, unlike the method of fundamental valuation that is applied when analyzing existing companies, it is not necessary to seek the help of sector stocks or sector experts, so it does not increase research-related costs (labor costs, etc.) The economy can be realized. As described above, according to the present invention, the larger the market is applied, the larger the profit can be expected at a lower cost, thereby providing a global hedge fund operation business model for the global market.

It is to be understood that the above-described embodiments are only illustrative of the constituent elements of the claims of the present invention, and do not limit the scope of the present invention, but rather should be construed as being included in the technical ideas throughout the specification of the present invention, Embodiments that include replaceable components as equivalents may be included within the scope of the present invention.

10:
20: transfer rate adjustment unit
30: financial variable estimation unit
100: asset management system
P: Investment Portfolio

Claims (11)

An error signal corresponding to a difference between a predetermined target financial variable and a total estimated financial variable of the portfolio is detected for the portfolio for asset management, and the individual estimated financial variable and each estimated financial variable for each item in the portfolio are detected. A control unit which outputs a control signal for determining an inclusion ratio for each item so that the error signal is minimized based on a change rate for a set time;
An incorporation ratio adjustment unit for converting an investment ratio of each item on the basis of the control signal and adjusting an incorporation ratio of each item in the portfolio based on the converted investment ratio; And
Based on the adjustment of the inclusion ratio adjustment of the inclusion ratio adjustment unit and information input from the market, the ratio of each item is re-detected from the portfolio whose characteristics are changed, and from the inclusion ratio of each item and the price and market index of each item newly measured A financial variable estimate that generates an individual estimated financial variable, estimates the rate of change of the individual estimated financial variable over time based on the individual estimated financial variable, and calculates an overall estimated financial variable of the portfolio based on the individual estimated financial variable Include a financial variable estimator that produces wealth;
Output of the financial variable estimating unit is input to the control unit asset management system through portfolio control, characterized in that the input. The method of claim 1,
A financial variable is an index indicating the sensitivity of a portfolio to a market change. An asset management system using portfolio control including at least one of a beta coefficient, a correlation coefficient, a volatility, and a delta coefficient according to the market. The method of claim 1,
The control unit comprises a Proportional Integral Derivational (PID) controller or a controller having a function of minimizing an error between the actual financial variable or the total estimated financial variable and the target financial variable. The method of claim 1,
The transfer ratio adjustment unit adjusts the transfer ratio of each item of the long position determined by the target financial variable and the control operation system so that the portfolio performance exceeds the reference index or the benchmark market index, or the portfolio performance Create an agenda index by converting the sum of the actual weights in each market to 100% to achieve positive returns on both ups and downs, and create individual agenda weights and target financial variables and control operations for the generated agenda indexes. The long position or the short position determined by calculating the difference between the specific gravity of each item of the long position determined by the system, and the short position or the short position corresponding to the long position or the short position, respectively, and the long position and the short position Incorporation ratio of each item is calculated so that the net position indicating the direction of Asset management system through portfolio control, characterized in that the adjustment. The method of claim 1,
The financial variable estimator comprises a Kalman filter or a filter having a function of removing noise of time series financial information. The method of claim 1,
The target financial variable is designated to move faster than the movement of the market when the rate of change of the entire market or the market index through the filtering in the financial variable estimator indicates a positive value, or at the time of the market index through the filtering. And a target financial variable setting unit for designating a target financial variable such that the rate of change for the negative value is equal to or slower than the movement of the market. In the asset management method of controlling the portfolio in a feedback control method through an asset management system having a control unit, an integrated ratio adjustment unit and a financial variable estimating unit,
Detecting an error signal corresponding to a difference between a predetermined target financial variable and a total estimated financial variable of the portfolio for the portfolio for asset management;
Outputting a control signal for determining an inclusion ratio for each item such that the error signal is minimized based on the individual estimated financial variable for each item in the portfolio and the rate of change of the individual estimated financial variable over a predetermined time period; ;
Converting the investment ratio of each item based on the control signal;
Performing at least one of buy, sell, short sell, and buyback for each stock so that the ratio of the stocks in the portfolio is adjusted based on the converted investment ratio;
The time of individual estimated financial variable and individual estimated financial variable at the inclusion ratio of each item is re-detected from the portfolio whose characteristics are changed by adjustment of the inclusion ratio of each item and information input from the market. Estimating a rate of change for; And
Calculating a total estimated financial variable of the portfolio based on the individual estimated financial variables and returning to the step of detecting the error signal for comparing the total estimated financial variable with the target financial variable of the next step.
Asset management method through portfolio control comprising a. The method of claim 7, wherein
The feedback may include filtering the target financial variable such that the sensitivity of the portfolio change is greater than the movement of the market when the rate of change for the predetermined time of the entire market or the market index is positive through filtering by the financial variable estimator. Or adjusting the target financial variable such that the sensitivity of the change in the portfolio is equal to or less than the movement of the market when the rate of change of the market index through the filtering indicates a negative value. How assets are managed through portfolio control. The method of claim 7, wherein
Outputting the control signal,
Calculating an actual weight portion of an index in a corresponding market index of each asset class of the portfolio;
Calculating the agenda weight ratio of each asset class in terms of 100% of the sum of the actual weight ratios; And
Recalculate the inclusion weight of each item in the portfolio so that the total estimated financial variable of the portfolio can follow the target financial variable, or the final inclusion of each asset item based on the difference between the recalculated inclusion weight and the agenda weight Step to adjust ratio
Asset management method through portfolio control comprising a. The method of claim 7, wherein
Estimating the rate of change of the individual estimated financial variable and the individual estimated financial variable over time includes estimating the individual estimated financial variable using a Kalman filter and estimating the rate of change of the individual estimated financial variable over time. How assets are managed through portfolio control. The method of claim 7, wherein
In the step of generating the individual estimated financial variable, when the financial variable is a beta coefficient, the return of the individual stocks of the portfolio is represented by Equation 11, the beta coefficient of the individual stocks, where the beta coefficient of the individual stocks Corresponding to the individual estimated financial variables-determined by Equation 12, the state variable (x) corresponding to the rate of change of the beta coefficient of the individual item over time is defined as shown in Equation 13, and When the state space function is set as in Equation 14, the state space function is estimated as in Equation 15, Equation 16, Equation 17 and Equation 18.
[Equation 11]

&Quot; (12) "

In Equations 11 and 12,

Respectively,
&Quot; (13) "

&Quot; (14) "

&Quot; (15) "

&Quot; (16) "

[Equation 17]

&Quot; (18) "

In Equations 15 to 18,

Respectively.

Images