Linear Multi-factor Model Research Articles

Statistical arbitrage in the stock markets by the means of multiple time horizons clustering

Nowadays, statistical arbitrage is one of the most attractive fields of study for researchers, and its applications are widely used also in the financial industry. In this work, we propose a new approach for statistical arbitrage based on clustering stocks according to their exposition on common risk factors. A linear multifactor model is exploited as theoretical background. The risk factors of such a model are extracted via Principal Component Analysis by looking at different time granularity. Furthermore, they are standardized to be handled by a feature selection technique, namely the Adaptive Lasso, whose aim is to find the factors that strongly drive each stock’s return. The assets are then clustered by using the information provided by the feature selection, and their exposition on each factor is deleted to obtain the statistical arbitrage. Finally, the Sequential Least SQuares Programming is used to determine the optimal weights to construct the portfolio. The proposed methodology is tested on the Italian, German, American, Japanese, Brazilian, and Indian Stock Markets. Its performances, evaluated through a Cross-Validation approach, are compared with three benchmarks to assess the robustness of our strategy.

Which Matters More for Migration: Evidence from Russian Regions’ Data

Many researchers of migration in Russia note deficiency of the works directed on modeling of migratory processes taking place in the country. This article is devoted to an assessment of impact power of socio-economic factors on interregional migration in Russia and to its comparison to recently found influence on this process of psychological characteristics of the accepting region population. As the aggregated at regional level estimates of intelligence, personal traits and characteristics of response style were calculated on the basis of results of online testing over 200 thousand respondents in 2012–19, indicators of socio-economic development of regions (n=16) are computed by averaging for the same time interval; the source of data - the Russian Federal Service of State Statistics (reference yearbooks “Regions of Russia” and the Uniform Interagency Information and Statistical System). In the analysis data of 78 subjects of the Russian Federation are included, but after association of capital agglomerations the number of regions was reduced to 76. It is established that the majority of the socio-economic indexes included in the analysis shows significant correlations with regional net migration coefficients. In general, the revealed pattern of correlations is agreed with results of other authors received on data of earlier periods of the analysis. Regression of net migration coefficients on socio-economic variables allowed calculating series of linear multifactor models. Best of these models accounted for about 44 % of a dependent variable variance. Earlier it was shown that the models calculated on the basis of limited number of aggregated psychological characteristics of the accepting region have the same level of accuracy. Moreover, addition of psychological variables to the best models based on socio-economic indexes provides essential increase of accuracy of prediction of regional net migration: the share of variance accounted for increases from 44 to 55–57 %. It is supposed that taking into account of psychological variables of the accepting region will create premises for more effective management of migratory processes.

ОСОБЛИВОСТІ ПРОГНОЗУВАННЯ ЧИСТОГО ДОХОДУ ВІД РЕАЛІЗАЦІЇ ПРОДУКЦІЇ (ТОВАРІВ, РОБІТ ТА ПОСЛУГ) В ДІЯЛЬНОСТІ ВИРОБНИЧИХ ПІДПРИЄМСТВ

In this article are considered the basic theoretical and methodological aspects of forecasting, the approaches of foreign and national scientists to the concept of «forecasting», as well as substantiated the necessity of forecasting in manufacturing companies' activities. It has been proposed a number of methods for forecasting the net income from the sale of products (goods, works and services) and has considered the features and conditions of their use. It has been summarized and substantiated the factors which have a direct impact on the net income from the sale of products, works and services of the company. These include the volume of intangible assets, the volume of fixed assets, the volume of current assets, the volume of net worth, the volume of own current capital, the annual inflation rate, and the exchange rate. There was formed the information-digital database (factor and result indicators) for the period from 2012 to 2022 on the basis of the materials of the Private Joint Stock Company «Food company "Podillya». By using the Microsoft Excel software product and specifically one of its «Data Analysis» services, the linear multi-factor model has been calculated. There are presented the results of correlation and regression analysis of the received statistical data for construction of economical and mathematical model and established density and quality of correlation interrelations between the selected factor quantities influencing the level of net income from sale of products (goods, works and services). The created forecast model made it possible to estimate internal and external factors which have an impact on the dynamics of net income from sales of products (goods, works and services). A forecast of the studied indicator in 2022 has been made by calculating the average growth coefficients of the factor indicators. The calculated correlation-regression model can be applied as a basis for the creation of management measures to implement an effective company development strategy in the future and to improve management decision-making approaches based on the results of economicmathematical modelling.

ЗАСТОСУВАННЯ МАТЕМАТИЧНОГО МОДЕЛЮВАННЯ ДО ПОБУДОВИ ЛІНІЙНОЇ БАГАТОФАКТОРНОЇ МОДЕЛІ ПЕРЕВЕЗЕНЬ ВАНТАЖІВ У МІЖНАРОДНОМУ СПОЛУЧЕННІ

The paper presents the problem of building multifactor models of international freight. 266 The object of the study is a multifactor model of international freight. The purpose of the work is to set the problem and build a linear multifactor model of international freight. Research methods include the main provisions of the theory of mathematical modeling, econometric methods of building models of real processes. The problem statement is considered and the linear multifactor model of international freight transport is constructed. The results obtained in the work can be used to predict the operation of a transport enterprise. KEY WORDS: LINEAR MODEL, MULTIFACTOR MODEL, INTERNATIONAL TRANSPORTATION, FREIGHT TRANSPORTATION, ECONOMETRIC METHODS

РОЗРОБЛЕННЯ МОДЕЛЕЙ ПРОГНОЗУВАННЯ ОБСЯГУ РЕАЛІЗАЦІЇ ПРОДУКЦІЇ СІЛЬСЬКОГОСПОДАРСЬКОГО ПІДПРИЄМСТВА

The financial and economic activities of economic entities in modern terms require constant plans comparison with real developments. Taking into account the peculiarities of the economic situation in the country, comparing the supply and demand, analysis of financial results, etc. The development of market relations increases the responsibility and independence of enterprises in making decisions about strategy and tactics for the future. That is why, while taking substantiating management decisions the application of techniques and methods of economic and mathematical modeling is relevant. One of the types of communication between the producer and the consumer is sales volume. It directly affects the amount of costs, profits and profitability of the enterprise. Therefore, the analysis and forecasting of such an indicator is important to increase the competitiveness of economic entities. Statistical data of one of the agricultural enterprises of Poltava region are analyzed, several influential factors are highlighted. It is proposed to perform a regression analysis, develop a linear multifactor model and build an adaptive model for forecasting the volume of sales of agricultural enterprises to select the optimal model for forecasting the studied indicator for the future. The author performed calculations on the obtained models. The analysis of the obtained results confirmed model’s adequacy, so they can be used for forecasting. As a result of comparison forecasts of sales volumes at the enterprise the adaptive model is recommended for practical application as it has the best quality of the forecast and the minimum sum of deviations squares. This model giving more weight to the latest statistics provides an optimistic forecast for the future. Thus, developing economic and mathematical forecasting models will provide an opportunity to improve the operational planning of production the certain types of products, as well as allow (after receiving new statistics) to quickly change the production monthly plans. Thus, conditions will be created to improve the quality of management, increase competitiveness and ensure the stabilization of the financial state at agricultural enterprises.

Моделирование показателя развития человеческого потенциала в Германии

The article is about a new model of the dependence of the state’s human resource index on a set of factors. The human resource index does not have a global assessment methodology, but it is very important for government management. One of the methods for the formal assessment of the characteristics of a large socio-economic system is the reasonable ranking of particular criteria. Purpose . The aim is a development of recommendations for increasing the indicators of human development index at the state level using only formal methods of mathematical modeling. Materials and methods . Particular criteria described the human development index are selected. Their ranking is determined, which ensures the greatest smoothness of the annual trend. A set of mathematical models is designed. They are estimated both by approximating the initial data and by the quality of post-forecast. The second-order regression-differential model is chosen as the working one, its advantages and disadvantages in compare with the linear multifactor model are shown. The influence of changes in controlled and uncontrolled factors on indicators of the human development index is studied. System management features are shown. Conclusion . Some recommendations are formulated based on short-term forecasts to improve the human development index.

Optimization of Stock Multifactor Model based on Machine Learning

In recent years, with the development of the financial market in China, in the background of information and big data, the quantitative products in the domestic market are gradually increasing. Multifactor model is an important stock selection model, its advantage is that it can synthesize a lot of information and get a stock selection result, which has a wide range of applications in the stock market. The purpose of this paper is to find some factors most related to the rate of return by establishing a multifactor model, and to select different weight factors to construct a stock selection model. The article is intended to select the stock combination to make it greater than or equal to the market index in the future, and to obtain the optimal benefit. Compared with the traditional linear multifactor model, machine learning algorithm can find more precise market signals through the nonlinear expression of factors, in order to obtain more robust excess returns.

PRICES FORECASTING IN THE PRIMARY HOUSING MARKET IN SAINT-PETERSBURG BASED ON REGRESSION- DIFFERENTIAL MODELING

The relevance of the study of the real estate prices dynamics in the city of St. Petersburg is shown. The criterion and factors influencing the object of modeling are determined. Normalized values of the criterion and factors are calculated. The mutual correlation of factors is investigated. A linear multifactor model and a model in the state space are built, it was proved that they cannot be used due to its poor predictive properties. A regression-differential model of prices changing for primary real estate in St. Petersburg is built, taking into account the mutual influence of factors, the possibilities of influence on its dynamics are revealed. Based on the linear multifactor model and the regression-differential model, the forecast of price changes for the next three years is made.

MODELING AND FORECASTING OF DOMESTIC AND INTERNATIONAL TOURISM IN TURKEY

An importance of tourism study and forecasting for the near future is proved. A criteria and factors affecting the object are choose from a set of publicly available annual series of selected private. The choice is justified. Normalized values of criteria and factors are calculated. A general criterion for assessing the quality of the object under study on the basis of private ones is compiled. The mutual correlation of factors is investigated. A linear multifactor model is built. It was proved that it cannot be used for prediction because of its poor predictive properties. A regression-differential model of the second order is constructed. An optimal combination of interpolation factors is selected. The forecast of factors and dynamics of changes in the object is calculated for the next three years. The influence of changes in controlled and uncontrollable factors on the object is investigated. The possibility of influencing an object under conditions of negative environmental effects is investigated. Recommendations for compensating of the negative impact are given as a decision support.

Investigating linear multi-factor models in asset pricing: considerable supplemental evidence

The literature has offered an interesting debate about whether the performance of Fama-French’s three-factor benchmark model is inadequate because it fails to pass some model specification tests and its R2 is not convincingly high in cross-sectional estimations. Previous studies have been quite limited, since they only focused on the time-series procedure with many models. We extend their work by providing a more robust investigation of the performance of several well-regarded pricing models in pooled portfolios and other portfolios sorted by new and important anomalies, using cross-sectional GMM tests for robustness. Finally, we find that, in addition to Fama and French’s five-factor model proposed in 1993, Fama-French’s three-factor model augmented by other factors usually outperforms Fama-French’s three-factor model across a significant proportion of different portfolios. In particular, Frazzini, Kabiller, and Pedersen’s model shows the best overall performance and consistency across different portfolios.

Portfolio performance of linear SDF models: an out-of-sample assessment

We evaluate linear stochastic discount factor models using an ex-post portfolio metric: the realized out-of-sample Sharpe ratio of mean–variance portfolios backed by alternative linear factor models. Using a sample of monthly US portfolio returns spanning the period 1968–2016, we find evidence that multifactor linear models have better empirical properties than the CAPM, not only when the cross-section of expected returns is evaluated in-sample, but also when they are used to inform one-month ahead portfolio selection. When we compare portfolios associated to multifactor models with mean–variance decisions implied by the single-factor CAPM, we document statistically significant differences in Sharpe ratios of up to 10 percent. Linear multifactor models that provide the best in-sample fit also yield the highest realized Sharpe ratios.

Portfolio Performance of Linear SDF Models: An Out-of-Sample Assessment

We evaluate linear stochastic discount factor models using an ex-post portfolio metric: the realized out-of-sample Sharpe ratio of mean-variance portfolios backed by alternative linear factor models. Using a sample of monthly US portfolio returns spanning the period 1968-2016, we find evidence that multifactor linear models have better empirical properties than the CAPM, not only when the cross-section of expected returns is evaluated in-sample, but also when they are used to inform one-month ahead portfolio selection. When we compare portfolios associated to multifactor models with mean-variance decisions implied by the single-factor CAPM, we document statistically significant differences in Sharpe ratios of up to 10 percent. Linear multifactor models that provide the best in-sample fit also yield the highest realized Sharpe ratios.

Prepotent response inhibition and reaction times in children with attention deficit/hyperactivity disorder from a Caribbean community.

Impairment in inhibitory control has been postulated as an underlying hallmark of attention deficit/hyperactivity disorder (ADHD), which can be utilized as a quantitative trait for genetic studies. Here, we evaluate whether inhibitory control, measured by simple automatized prepotent response (PR) inhibition variables, is a robust discriminant function for the diagnosis of ADHD in children and can be used as an endophenotype for future genetic studies. One hundred fifty-two school children (30.9% female, 67.8% with ADHD) were recruited. The ADHD checklist was used as the screening tool, whilst the DSM-IV Mini International Neuropsychiatry Interview, neurologic interview and neurologic examination, and the WISC III FSIQ test were administered as the gold standard procedure to assert ADHD diagnosis. A Go/No-Go task using a naturalistic and automatized visual signal was administered. A linear multifactor model (MANOVA) was fitted to compare groups including ADHD status, age, and gender as multiple independent factors. Linear discriminant analysis and the receiver operating characteristic curve were used to assess the predictive performance of PR inhibition variables for ADHD diagnosis. We found that four variables of prepotent response reaction time- and prepotent response inhibition established statistically significant differences between children with and without ADHD. Furthermore, these variables generated a strong discriminant function with a total classification capability of 73, 84% specificity, 68% sensitivity, and 90% positive predictive value for ADHD diagnosis, which support reaction times as a candidate endophenotype that could potentially be used in future ADHD genetic research.

Orthogonalized factors and systematic risk decomposition

Abstract In the context of linear multi-factor models, this study proposes an egalitarian, optimal and unique procedure to find orthogonalized factors, which also facilitates the decomposition of the coefficient of determination. Importantly, the new risk factors may diverge significantly from the original ones. The decomposition of risk allows one to explicitly examine the impact of individual factors on the return variation of risky assets, which provides discriminative power for factor selection. The procedure is experimentally robust even for small samples. Empirically we find that even though, on average, approximately eighty (sixty-five) percent of style (industry) portfolios’ volatility is explained by the market and size factors, other factors such as value , momentum and contrarian still play an important role for certain portfolios. The components of systematic risk, while dynamic over time, generally exhibit negative correlation between market , on one side, and size , value , momentum and contrarian , on the other side.

Stock Return Prediction and Anomaly Detection by Regression Trees

Noisy markets need extensive descriptions that are noisy themselves, such as deep regression trees that capture many data-local nonlinear anomalies and that do not require arbitrary weighting schemes like traditional linear multifactor models often do. Simple tools allow extraction of general and stock specific insightful information from such regression trees and facilitate factor selection. Large tree-models show excellent out-of-sample predictive power, that we measure and discuss extensively. The method is described by a seven-factor example model for the worldwide industrial sector that delivers 20% quintile-separation excess return out-of-sample and that reveals for instance how following price falls analysts react belatedly and investors overreact and thus create reversal outperformance opportunities.

A comparison of linear and nonlinear statistical techniques in performance attribution

Performance attribution is usually conducted under the linear framework of multifactor models. Although commonly used by practitioners in finance, linear multifactor models are known to be less than satisfactory in many situations. After a brief survey of nonlinear methods, nonlinear statistical techniques are applied to performance attribution of a portfolio constructed from a fixed universe of stocks using factors derived from some commonly used cross sectional linear multifactor models. By rebalancing this portfolio monthly, the cumulative returns for procedures based on standard linear multifactor model and three nonlinear techniques-model selection, additive models, and neural networks-are calculated and compared. It is found that the first two nonlinear techniques, especially in combination, outperform the standard linear model. The results in the neural-network case are inconclusive because of the great variety of possible models. Although these methods are more complicated and may require some tuning, toolboxes are developed and suggestions on calibration are proposed. This paper demonstrates the usefulness of modern nonlinear statistical techniques in performance attribution.

Intertemporal Variation in the Performance of Hedge Funds Employing a Contingent-Claim-Based Benchmark

Since it is well known that hedge fund returns exhibit non-linear option-like exposures to standard asset classes, traditional linear factor models offer limited help in capturing the risk-return tradeoffs offered by hedge funds. This paper employs a combination of passive buy-and-hold strategies and option-based strategies to characterize the risks of different hedge fund strategies. Although, in practice, these hedge funds can follow a myriad of dynamic trading strategies, we find that adding a few simple option writing/buying strategies to the linear multi-factor model enables us to capture a significant proportion of the variation in the hedge fund returns over time. We verify the ability of our approach to capture important hedge fund risk exposures by conducting out-of-sample analysis. Our approach can provide valuable insights into the nature of risks involved in investing in different hedge fund strategies.