Professional Documents
Culture Documents
Optimization of Investment Portfolio Management
Optimization of Investment Portfolio Management
Abstract
The task of creating an investment portfolio by a financial institution is considered. Funds for
creating a portfolio are taken from two sources: enterprise's equity funds and borrowed funds.
Optimization of the created portfolio is performed. A portfolio of maximum efficiency was obtained
with restriction on the measure of risk, which is specified in the form of a VaR indicator. Using
optimization portfolio data, a model of portfolio asset management is being built. Using the
Pontryagin maximum principle, optimal strategies of its participants are determined. The optimal
function of managing the investment portfolio in the form of a share of the income received is found.
Numerical results of optimal management of investments in a financial portfolio from the financial
institution as well as from the creditor are presented.
doi: 10.5937/sjm14-16806
xxx V. Oliinyk / SJM 14 (2) (2019) xxx - xxx
account financial interests of all its to manage investors' financing. The
participants. The main tasks to be objective in this formulation is reduced to a
considered in this article are as follows: differential game with no opposing interests
construction of an investment portfolio based regarding the financial institution
on the Markowitz theory; task formulation (enterprise) and external investors
for the optimal management of financial (Germeier, 1976).
assets using the Pontryagin maximum
principle; formulation of interests for all the
participants of the system; use of the utility 2. LiTERATuRE REViEW
function to construct the objective function;
finding the optimal control function (OCF); Many authors studied the formation of an
obtaining the results of numerical optimization portfolio, which would consist
experiments. of risk and risk-free financial instruments.
A new investment portfolio is assumed to One of the main criteria for building a
be created. The portfolio suggests every portfolio is its risk index. To build
participant to have his own quota, which is optimization models and risk analysis, Shalit
maintained throughout the entire period of & Yitzhaki (1984) suggested using the
the project's existence. To create an Mean-Gini (MG) method, which is simple in
investment portfolio of maximum efficiency, its practical application.
the Markowitz model (1952) is used. In his In their work, Kalayci et al. (2019)
work, Markowitz introduces the concept of analyze the results of research on the
risk in the form of standard deviation and formation of an optimal securities portfolio.
forms a portfolio considering this index. In Much attention is paid to algorithms and
this paper, it is proposed to form an methodologies based on the use of computer
investment portfolio, where Value-at-risk technology. A review of the previously
(VaR) is used as an indicator of risk. It constructed portfolios that improve the
should be noted that this indicator reflects standard portfolio in a deterministic
the amount of losses for a certain period in formulation was made. Based on a
money terms. comprehensive review, a future research area
Having formed the optimal portfolio of is proposed in this area.
maximum efficiency with regard to the Bender et al. (2019) propose building
restrictions on possible financial losses, the portfolios of factors based on average values
financial development of this project should of variance. In their work, they analyze
be implemented for some time. Depending portfolios based on alternative weighting
upon the funds invested and the financial schemes. It is shown that the portfolios of
interests of each of the participants, the factors describe the information more fully
stages of attracting monetary resources due to the explicit use of input factors.
should be made up. Thus, an initial project is Danko & Šoltés (2018) in their work
being developed. One of the methods of proposes to form an investment portfolio
optimal management of the acquired based on Markowitz theory in combination
investment project is the one based on the with graph theory. Testing of the resulting
Pontryagin maximum principle (Pontryagin portfolios was carried out using simulation
et al., 1962). It is proposed to use this method analysis. The results can be used by
V. Oliinyk / SJM 14 (2) (2019) xxx - xxx xxx
individual investors in the formation of the consider portfolio optimization based on an
optimization portfolio. empirical likelihood estimation method and a
Zhou et al. (2019) consider the optimal criterion for making a decision on stochastic
investment portfolio at risk of a jump over dominance. This approach is implemented in
time. It is shown that for the formation of the two stages: the necessary probabilities are
optimization portfolio important indicators found using convex programming; with the
are: the initial intensity of the jump, the use of linear programming formed an
dynamics and stability of the jump. Formed optimization portfolio. The resulting
investment portfolio provides good out-of- portfolio gives good results for the out-of-
sample performance. sample impulse strategy. Grinblatt and
Calvo et al. (2018) propose to form an Saxena (2018) propose to improve the
investment portfolio based not only on accuracy of determining the pricing of
historical data. For this purpose, it is imitating factors by obtaining a single
proposed to use the concept of value portfolio by combining basic portfolios. This
information. This indicator reflects investor approach can also be used for portfolios with
confidence in his non-historical data. The anomalies.
level of risk outside the historical interval is Since the mid-1990s, researchers began
proposed to be adjusted with the help of using the Value-at-Risk (VaR) indicator as a
diversification restrictions. Lester (2019) market risk. Holton (2003) publishes a book
builds a portfolio based on the theory of about the use of the VaR indicator at the real
factor investment. A comparison is made trading floors. In his publication, he shows
between a set of single-factor investment how to design and implement this indicator
portfolios and a single integrated portfolio. It in real time mode.
is shown that the integrated portfolio more Basak and Shapiro (2001) analyzed the
accurately analytically predicts profitability use of VaR criteria by managers while
and risk in subsequent investment periods. forming the dynamic optimization portfolios.
The application of portfolio theory in They found out that risk managers using VaR
insurance is presented in Oliynyk (2015). An opted for a risky portfolio management
algorithm for obtaining an optimization strategy and, therefore, their activities could
portfolio of the insurance company is lead to significant financial losses. To
presented. For real data, the optimal minimize these losses, they proposed an
composition of the investment insurance alternative risk management strategy based
portfolio of minimum risk by type of on the expectation of losses. Zhang et al.
insurance was obtained. As limitations of the (2019) in their article for the analysis of
task, the main indicators of the insurance maximum loss, they propose to use the
company’s financial activities are used. double-VaR indicator. This indicator
Uhl and Rohner (2018) proposes to summarizes the obtained values of the
introduce a compensation portfolio that standard one-dimensional VaR. The two-
optimizes asset allocation and related risks. dimensional income-risk confidence domain
The proposed approach allows for the is analyzed for two models. Francq and
optimization of portfolio parameters using Zakoïan (2018) considers a dynamic model
both the theory of behavioral portfolio and of changes in the composition of assets and
modern portfolio theory. Post et al. (2018) conduct a joint assessment of market and
xxx V. Oliinyk / SJM 14 (2) (2019) xxx - xxx
appraisal risks in the formed portfolio. Under bank - maximizing the exponential utility
the conditions of elliptic distribution, the function. An explicit solution of optimal
value of VaR acts as a measure of risk. We control problems is given. Bilbao-Terol et al.
also study filtered historical modeling, a (2016) propose a mathematical model based
research method that allows for more general on targeted programming, which covers the
distribution conditions. A portfolio with theory of investment portfolio with mental
minimal variance is explored in more detail. accounting and responsible investment. The
The advantages and disadvantages of the share participation of each investor is
considered approaches are analyzed. proposed to be assessed by an expert method
Burdorf and Van Vuuren (2018) in their taking into account fuzzy rules. Mei and
study analyzes the market behavior of risk Nogales (2018) explores the behavior of a
measures VaR and Expected Shortfall (ES). multi-period investor when choosing a
Considering the stock portfolio, it is shown portfolio. The optimality criterion is its
that finding these indicators by the Monte maximum utility of intermediate
Carlo method and the dispersion-covariance consumption, which depends on risky assets,
method is preferable to methods based on costs and predictable profitability. An
historical data. Rockafellar and Uryasev optimal consumption algorithm is proposed
(2002) suggested using conditional value-at- based on the proposed suboptimal plans.
risk (CVaR) as a measure of risk. This For successful functioning of a financial
criterion makes it possible to obtain risk enterprise, it is necessary to form an
assessments that do not fully reflect the VaR investment portfolio as well as to develop
indicator. When applying this indicator, a and manage it in the best possible way.
modern mathematical apparatus should be Van Gelderen et al. (2019) analyze the
used and computational experiments carried management strategy of a mutual investment
out. fund using the factor investment theory. On
To model investors’ behavior, the utility the basis of the conducted studies, it was
function could be used. This function must shown that the buy-and-hold strategy
meet certain requirements and it reflects the together with the choice of a multifactorial
attitude to the risk of the person making the manager is more profitable than the tactics of
decision. Blanchett and Ratner (2015) active fund management. Del Guercioa
introduce a new utility function that includes et al. (2018) study the effectiveness of
the risk of a decline. This function is used portfolio management, consisting of three
when forming an optimal portfolio for an types of accounts: hedge funds, mutual funds
investor focused on income generation. and separate accounts. In their studies, they
Maheshwari and Sarantsev (2018) is confirm the hypothesis of a conflict of
considering an optimal risk management interest with the parallel management of the
model. The investors are private entities that created portfolio.
borrowed money from the central bank at a Li et al. (2018) consider the problem of
controlled interest rate. It is assumed that describing the profitability of assets using
each of the participants of the investment non-parametric programming. An approach
project seeks to improve their well-being: a based on the Fama – French three-factor
private entity through the maximization of model is proposed: cost, size and market
the expected logarithmic wealth; central factor. For each factor, transformations are
V. Oliinyk / SJM 14 (2) (2019) xxx - xxx xxx
constructed, the necessity of which is The article Oliinyk (2017) discusses the
verified by a generalized criterion of the problem of optimal control of the
maximum likelihood hypothesis. The distribution of gross product when
proposed approach is tested on real data and considering a single-productive economic
samples of finite sizes. Simonian and Wu system. The control function is the value of
(2019) learn replication hedge funds when the share of the gross product, aimed at non-
investing. For accurate replication, it is productive consumption. Analyzes the
proposed to use the ridge regression method. switching point of investment at which the
This method allows to obtain a generalized company will receive the maximum income.
model that most adequately analyzes the To solve this problem, the Pontryagin
risks and improves the basic factorial models maximum method is used.
of hedge fund replication based on ordinary Hartl et al. (1995) considered various
least squares (OLS) regression. forms of the Pontryagin maximum principle
García-Melón et al. (2016) consider a for optimal management issues with
financial asset with a social responsibility restrictions on state variables. On the
(SR). SR involves finding the various criteria illustrative examples, they showed the
and their weights in the asset under application of these optimal forms for
investigation. When finding these solving managerial tasks.
parameters, it is necessary to take into Aseev et al. (2005) studied the model of
account the preferences of each investor. In endogenous growth of two countries, based
this article, the authors suggest using the on the Pontryagin maximum principle. It is
Analytic Hierarchy Process to determine shown that one of the countries develops due
weights. to the absorption of some knowledge
One of the methods for optimal control of obtained in the leading country. The work of
dynamical systems is the one based on the Aseev and Kryazhimskii (2007) is dedicated
Pontryagin maximum principle. Many to the theoretical aspects of the Pontryagin
authors used this principle in their scientific maximum principle for solving optimal
works. Koopmans (1967) reviewed the management tasks arising in the economy
results of economic growth models based on when studying the processes of economic
the Pontryagin maximum principle. He also growth. Aseev (2014) considered application
examined the obstacles that arise when of the Pontryagin maximum principle for a
applying this method and suggested class of optimal management tasks with an
directions for further research. Shell (1969) infinite horizon that appears when studying
in his work considered the application of the the economic growth processes.
Pontryagin maximum principle in the Krasovskii and Tarasyev (2008)
economy. considered the task of optimal management
Kamien and Schwartz (1971) showed that with the functional given by the improper
the Pontryagin method could be applied to integral. This task arises when considering
the problems arising in economic analysis. models of economic growth. The proposed
They have proved the theorem concerning algorithm, based on the Pontryagin
the properties of the state variables of the maximum principle, makes it possible to
system, which makes it possible to obtain the obtain optimal trajectories for real
optimal control solution for a certain class of macroeconomic data.
economic issues.
xxx V. Oliinyk / SJM 14 (2) (2019) xxx - xxx
Option k (t ) d (t ) E p (t ) δ (t ) λ
0 -1.474
1 0.6 0 0.25 0 -0.172
1 -0.042
2 0.6 0.3 0.25 0 -0.213
0.2 -0.264
3 0.6 0 0.25 0 -0.172
0.3 -0.114
0 -0.213
4 0.6 0.3 0.25 0.05 -0.197
0.1 -0.175
0 -0.341
5 0.0777t + 0.0223 0.0333t – 0.0333 0.0222t + 0.0778 0.05 -0.287
0.1 -0.235
Figure 2. The distribution of the optimal control function (Option 2)
Mathematics, 257 (1), 1-255. Del Guercioa, D., Genc, E., & Tran, H.
Aseev, S., Hutschenreiter, G., (2018). Playing favorites: Conflicts of
Kryazhimskiy, A., & Lysenko, A. (2005). A interest in mutual fund management. Journal
dynamic model of optimal investment in of Financial Economics, 128 (3), 535-557.
research and development with international Francq, C., & Zakoïan, J.-M. (2018).
knowledge spillovers. Mathematical and Estimation risk for the VaR of portfolios
Computer Modeling of Dynamical Systems, driven by semi-parametric multivariate
11 (2), 125-133. models. Journal of Econometrics, 205 (2),
Basak, S., & Shapiro, A. (2001). Value-at- 381-401.
Risk Based Risk Management: Optimal García-Melón, M., Pérez-Gladish, B.,
Policies and Asset Prices. Review of Gómez-Navarro, T., & Mendez-Rodriguez,
Financial Studies, 14 (2), 371-405. P. (2016). Assessing mutual funds’ corporate
Bender, J., Blackbur, T., & Sun, X. social responsibility: a multistakeholder-
(2019). Clash of the Titans: Factor Portfolios AHP based methodology. Annals of
versus Alternative Weighting Schemes. The Operations Research, 244 (2), 475-503.
Journal of Portfolio Management Germeier, Y.V. (1976). Games with non-
Quantitative, 45 (3), 38-49. opposite interests (in Russian). Nauka,
Bilbao-Terol, A., Arenas-Parra, M., Moscow.
Cañal-Fernández, V., & Bilbao-Terol, C. Grinblatt, M., & Saxena, K. (2018). When
(2016). Multi-criteria decision making for Factors Do Not Span Their Basis Portfolios.
choosing socially responsible investment Journal of Financial and Quantitative
within a behavioral portfolio theory Analysis, 53 (6), 2335-2354.
framework: a new way of investing into a Hartl, R.F., Sethi, S.P., & Vickson, R.G.
crisis environment. Annals of Operations (1995). A Survey of the Maximum Principles
Research, 247 (2), 549-580. for Optimal Control Problems with State
Blanchett, D., & Ratner, H. (2015). Constraints. SIAM Review, 37 (2), 181-218.
Building Efficient Income Portfolios. The Holton, G.A. (2003). Value-at-Risk:
Journal of Portfolio Management, 41 (3), theory and practice. Academic Press. San
117-125. Diego. USA
Burdorf, T., & Van Vuuren, G. (2018). An Kalayci, C.B., Ertenlice, O., & Akbay,
evaluation and comparison of Value at Risk M.A. (2019). A comprehensive review of
and Expected Shortfall. Investment deterministic models and applications for
Management and Financial Innovations, 15 mean-variance portfolio optimization.
(4), 17-34. Expert Systems with Applications, 125 (1),
Calvo, C., Ivorra, C., & Liern, V. (2018). 345-368.
Controlling risk through diversification in Kamien, M.I., & Schwartz, N.L. (1971).
portfolio selection with non-historical Sufficient conditions in optimal control
information. Journal of the Operational theory. Journal of Economic Theory, 3 (2),
Research Society, 69 (10), 1543-1548. 207-214.
Danko, J., & Šoltés, V. (2018). Portfolio Koopmans, T.C. (1967). Objectives,
creation using graph characteristics. Constraints, and Outcomes in Optimal
Investment Management and Financial Growth Models. Econometrica, 35 (1), 1-15.
Innovations, 15 (1), 180-189. Krasovskii, A.A., & Tarasyev, A.M.
V. Oliinyk / SJM 14 (2) (2019) xxx - xxx xxx
(2008). Conjugation of Hamiltonian Systems (2018). Portfolio optimization based on
in Optimal Control Problems. Proceedings of stochastic dominance and empirical
the 17th IFAC World Congress, South Korea, likelihood. Journal of Econometrics, 206 (1),
7784-7789. 167-186.
Lester, A. (2019). On the Theory and Rockafellar, R., & Uryasev, S. (2002).
Practice of Multifactor Portfolio. The Journal Conditional Value-at-Risk for General Loss
of Portfolio Management Quantitative, 45 Distributions. Journal of Banking & Finance,
(3), 87-100. 26 (7), 1443-1471.
Li, J., Zhang, W., & Kong, E. (2018). Shalit, H., & Yitzhaki, S. (1984). Mean-
Factor models for asset returns based on Gini, portfolio theory, and the pricing of
transformed factors. Journal of risky assets. The Journal of Finance, 39 (5),
Econometrics, 207 (2), 432-448. 1449-1468.
Maheshwari, A., & Sarantsev, A. (2018). Shell, K. (1969). Application of the
Modeling Financial System with Interbank Pontryagin's maximum principle to
Flows, Borrowing, and Investing. Risks, 6 economics. In Lecture Notes in Operations
(4), 131. Research and Mathematical Economics, 11.
Markowitz, H. (1952). Portfolio selection. Mathematical system theory and economics,
The Journal of Finance, 7 (1), 77-91. 1, 241-292, Berlin, Springer.
Mas-Colell, A., Whinston, M.D., & Simonian, J., & Wu, C. (2019). Factors in
Green, J.R. (1995). Microeconomic Theory. Time: Fine-Tuning Hedge Fund Replication.
Oxford University Press, Oxford. The Journal of Portfolio Management
Mei, X., & Nogales, F.J. (2018). Portfolio Quantitative, 45 (3), 159-164.
Selection with Proportional Transaction Uhl, M.W., & Rohner, P. (2018). The
Costs and Predictability. Journal of Banking compensation portfolio. Finance Research
& Finance, 94(C), 131-151. Letters, 27, 60-64.
Oliinyk, V. (2017). Optimal Management of Van Gelderen, E., Huij, J., & Kyosev, G.
the Enterprise's Financial Flows. Journal of (2019). Factor Investing from Concept to
Advanced Research in Law and Economics, Implementation. The Journal of Portfolio
8 (6), 1875-1883. Management Quantitative, 45 (3) 125-140.
Oliynyk, V. (2015). Modeling of the Zhang, W., Zhang, S., & Zhao, P. (2019).
optimal structure of insurance portfolio. On Double Value at Risk. Risks, 7 (1), 31.
Problems and Perspectives in Management, Zhou, C., Wu, C., & Wang, Y. (2019).
13(2), 230-234. Dynamic portfolio allocation with time-
Pavlov O.V. (2004). Dynamic models of varying jump risk. Journal of Empirical
interaction of participants in corporate Finance, 50, 113-124.
systems (in Russian). Management of large
systems: collection of paers, 8, 157-175.
Pontryagin, L.S, Boltyanskii, V.G,
Gamkrelidze, R.V. & Mishchenko, E.F.
(1962). The mathematical theory of optimal
processes. Intersciene Publishers John
Wiley&Sons, Inc. New York-London.
Post, T., Karabatı, S., & Arvanitis, S.