نشریه تحقیقات مالی
پیاپی 77 (بهار 1404)

The objective of the present research is the optimization of investment portfolios, considering all significant criteria for investors in a fuzzy environment, and taking into account various levels of investor expectations for each criterion based on their risk preferences. On one hand, the utilization of fuzzy logic in modeling this problem can enhance its alignment with real-world conditions by accommodating uncertainty in input data. On the other hand, the proposed model for portfolio selection incorporates not only risk and return as key factors for investors but also considers other important factors. These factors include short-term and long-term returns, liquidity, maximum and minimum investment ratios in each asset, dividend distribution, and cardinality constraint (the number of assets within the portfolio). Another objective of this research is to present an innovative model compared to existing ones by incorporating a logistic fuzzy membership function to model various levels of investor expectations. This enables the formation of investment portfolios based on the priorities of investors with different risk appetites regarding different criteria.

The method of conducting this research involves initially addressing the optimization modeling of a nonlinear multi-objective problem in a fuzzy environment, considering all crucial factors for investors. Subsequently, employing quantitative methods and the foundations of fuzzy logic, we transform the problem into a single-objective linear problem, making it amenable to solution using conventional optimization methods and software. Ultimately, utilizing data from the 50 most active companies on the Tehran Stock Exchange (TSE) market, we implement the model and analyze the results.

The research results indicate that the proposed model yields a return more than twice that of the index of the 50 most active companies on the Tehran Stock Exchange (TSE) for both aggressive and conservative investors. Additionally, the model, by incorporating a logistic-shaped membership function for various problem objectives, can be customized for investors with aggressive (risk-tolerant) or conservative (risk-averse) strategies. This customization is attributed to a parameter that determines the shape of the membership function in logistic functions, allowing the prioritization of different factors such as risk, short-term and long-term returns, liquidity, or dividend distribution for investors with varying levels of risk tolerance.

The utilization of logistic-shaped membership functions in a fuzzy environment, along with diverse criteria, can personalize the investment portfolio selection model for investors with different characteristics. This customization enables investors with various risk tolerances to construct a portfolio according to their priorities. This adaptability significantly enhances the practical applicability of the portfolio selection problem in the real world. Furthermore, employing computational methods and fuzzy principles allows the transformation of a nonlinear multi-objective problem into a single-objective linear model, facilitating its implementation and resolution.

The aim of predicting profit changes is to create awareness for investors, financial analysts, managers, stock market officials, creditors, and other users to judge the business unit, make decisions about buying or selling stocks, or granting or denying loans and credits. The goal of this research is to evaluate the performance and compare the accuracy of machine learning models and statistical models in predicting the direction of changes in three profit components including net profit (loss), gross profit (loss), and operating profit (loss).

In this research, using the financial information of 139 manufacturing companies listed on the Tehran Stock Exchange over a 15-year period, from 2008 to 2022, and employing 25 machine learning models and 10 statistical models, the efficiency of machine learning models and statistical models in predicting the direction of changes in profit components including net profit (loss), gross profit (loss), and operating profit (loss) has been compared. In the present study, Excel software was used for data sorting, Eviews software for extracting descriptive statistics, and data mining software SPSS Modeler and Rapidminer for predicting profit changes. The performance of machine learning models was evaluated using two criteria: accuracy (predictive accuracy of the model) and AUC (area under the curve), and the performance of statistical models was evaluated only by the accuracy criterion. Finally, in order to select the model with the best performance for predicting the direction of changes in net profit (loss), gross profit (loss), and operating profit (loss), the best model among the machine learning models was chosen using the ROC curve.

After calculating the average predictive accuracy of machine learning and statistical models, it was found that the average predictive accuracy of machine learning models for dependent variables including the percentage of changes in net profit (loss), percentage of changes in gross profit (loss), and percentage of changes in operating profit (loss) ranges from 83% to 93%. It was also found that the average predictive accuracy of statistical models for all three profit components varies from 76% to 83%. After confirming the non-normality of the average accuracy of machine learning and statistical models for profit components using the Kolmogorov-Smirnov test, the non-parametric Mann-Whitney U test was used to compare the predictive accuracy of machine learning models and statistical models in predicting the direction of changes in profit components.

The results of the research hypotheses test indicate the high efficiency of machine learning models in predicting the direction of changes in net profit (loss), gross profit (loss), and operating profit (loss), compared to statistical models. The ROC curve results indicate that the decision tree model achieved a predictive accuracy of 100% in forecasting the direction of changes in net profit (loss) and 99.38% accuracy in predicting the direction of changes in gross profit (loss). Additionally, the rule-based inference model demonstrated a predictive accuracy of 86.76% for forecasting the direction of changes in operating profit (loss). These models exhibited the best performance and were selected as the optimal models.

In recent years, the increasing demand for energy has highlighted the importance of energy in production. As a strategic energy source, natural gas is a vital component of the production process for goods and services. On the other hand, financial markets are crucial to any country's economy, and stock exchanges are a key component of these markets. Various indices calculated on stock exchanges indicate the return and overall trend of stock prices in the entire market and specific industries. The fluctuation of each index is influenced by changes in the stock prices of companies included in the calculation. Many factors contribute to fluctuations in companies' stock prices, including the price of natural gas, which is used by various industries in their production processes. Changes in gas prices affect companies' profit margins, leading to changes in their stock prices. Ultimately, these changes impact the overall return of the capital market. Given the significance of this topic, this research examines the short-term and long-term effects of changes in natural gas prices on the return of the Iranian stock market.

This research is classified as applied research, employing quantitative methodology. To analyze the monthly data spanning from April 2009 to March 2022, this study utilized the Vector Autoregression (VAR) method and the Johansen Cointegration Test.

The results confirm a long-term relationship between natural gas price changes and selected capital market indices. The Johansen Cointegration Test indicates that changes in gas prices have a negative impact on stock market indices, OTC, chemical industries, industry, cement, and petroleum products in both the short- and long-term. Furthermore, the results indicate that exchange rate changes have a positive and significant effect on capital market indices in both the short- and long-term. Finally, the test results show that oil price changes have a negative impact on indices in the short-term, but in the long-term, changes in indices will move in the same direction as oil price changes.

Considering the significant impact of gas prices on capital market fluctuations, it is recommended that the government regulate gas prices to minimize their negative effects on capital market indices. To achieve this, sudden decisions regarding gas prices should be avoided. Instead of making annual budget decisions, a dynamic gas pricing formula should be designed, allowing the price of gas offered to industries to fluctuate in response to global gas price changes. This would provide investors with the stable, long-term economic conditions they require. To promote diversification of energy consumption, the government can implement policies to increase investment in renewable and emerging energy sectors, encouraging companies to switch to alternative energy sources. This would reduce industries' dependence on gas and mitigate the impact of gas price changes on the stock market.

Predicting the future trends in financial markets stands as a critical task for both investors and researchers, given its pivotal role in enabling well-informed decision-making processes and effective risk management strategies. Nevertheless, the realm of stock market dynamics is fraught with inherent complexities and uncertainties, posing a formidable challenge when it comes to achieving accurate predictions. A wide array of predictive modeling techniques have been meticulously investigated, spanning from conventional statistical methodologies to more sophisticated machine learning algorithms. The primary focus of this research endeavor revolves around the predictive analysis of the Tehran Stock Exchange (TSE) Composite Index, wherein a novel hybrid neural network framework is employed. This approach seamlessly integrates multiscale temporal features, with the ultimate objective of bolstering prediction precision and offering profound insights into prevailing market trends and dynamics.

The hybrid neural network architecture that has been put forward integrates the unique capabilities of convolutional neural networks (CNNs) in the realm of feature extraction with the effectiveness of long short-term memory (LSTM) networks in capturing temporal dependencies. The dataset used in this study consists of daily historical data pertaining to the TSE Composite Index, covering a substantial period from the year 1998 to 2022, which has been meticulously gathered, preprocessed, and subsequently partitioned into distinct sets for training and validation purposes. Within the framework of this hybrid neural network model, a sophisticated approach is adopted to harness multiscale temporal features derived from the input data, enabling the generation of highly accurate predictions regarding the future trends of the index. Moreover, to further enhance the performance and resilience of the model, sophisticated feature engineering methodologies are implemented to optimize its overall functionality.

The results of the study reveal that while the hybrid neural network model, integrating CNN and LSTM components, demonstrates promising capabilities in predicting the TSE Composite Index, its accuracy falls short compared to competing models, particularly at weekly and monthly time scales. Conversely, the standalone CNN model exhibits superior performance, yielding more accurate predictions of the index's movements. These findings challenge the prevailing notion regarding the efficacy of hybrid neural network models in financial market prediction, highlighting the importance of evaluating alternative modeling approaches based on their specific strengths and limitations.

Despite the potential of hybrid neural network models, as demonstrated in previous research, the findings of this study suggest that simpler neural network architectures, such as CNNs, may offer better prediction performance in certain scenarios. To address the limitations identified, future research endeavors could explore alternative model configurations, ensemble methods, or hybrid architectures that combine the strengths of different predictive models. Additionally, incorporating additional market indicators and exploring intraday data sources could further enhance prediction accuracy and robustness. This abstract encapsulates the key findings and implications of the research, providing valuable insights for investors, researchers, and practitioners in the field of financial market prediction.

Financial literacy is widely recognized as a crucial life skill, with individuals in developed countries expected to attain a desirable level of financial competence. This significance is reflected in the fact that more than 60 countries have developed national policies and strategies to promote financial literacy, incorporating specific programs into their agendas. Iran is no exception. Enhancing financial literacy is particularly important given the country's distinct economic, social, political, and cultural conditions. Improving individuals' financial awareness, knowledge, and skills reduces financial errors, strengthens sound financial decision-making, boosts confidence, and aids in achieving financial security and well-being. Furthermore, it can contribute to macroeconomic and social improvements, ensure economic stability, and foster greater public alignment with financial and economic policies. Therefore, it is essential to integrate this issue into the agendas of higher-level policymaking institutions and develop policies to promote financial literacy that are tailored to Iran's unique Islamic-Iranian culture.

This study was conducted by collecting policy documents and national strategies from other countries, as well as through semi-structured interviews with experts, specialists, and scholars in financial management, financial literacy, and policymaking. It employs qualitative thematic analysis to explore the main question: What policy framework is needed to promote financial literacy in Iran?

Research findings reveal that the policy framework for enhancing financial literacy in Iran can be structured around four primary themes, each comprising several subthemes. The domains include: 1. Policy prerequisites: (including legal, financial, executive, and cultural prerequisites); 2. Policy objectives (including educational, economic, and socio-political goals); 3. Policy requirements (including content development, audience analysis, expert cultivation, and evaluation measures); and 4. Policy tools (including traditional, classic, innovative, and supportive-motivational mechanisms). This framework can serve as a blueprint for policymaking in the field of financial literacy enhancement in Iran.

Financial literacy enhancement programs in any country must be designed in accordance with the country's unique circumstances. Given Iran's unique context, the proposed framework—grounded in international models and domestic expert insights—may guide policymakers in developing effective, localized policies in this domain.

This study aims to develop threshold asset pricing models to enhance the performance of common multi-factor models. Over the past thirty years, asset pricing models have evolved by incorporating pricing anomalies as new factors that previous models could not explain, leading to the introduction of hundreds of such factors. This proliferation underscores the importance of the “parsimony” principle, which advocates for models with minimal factors and maximal explanatory power. Given the abundance of potential variables influencing asset returns, it is essential to seek models that balance simplicity and effectiveness. In line with this principle, this study proposes threshold asset pricing models where the proposed factors are applied selectively to certain companies rather than universally. Through cross-sectional tests and threshold cross-sectional regression analyses, the research examines the threshold effect of variables such as credit risk on expected returns, using the debt ratio as a threshold variable. It is anticipated that the impact of credit risk or financial distress on expected returns will differ across various levels of indebtedness. If the threshold effect of the debt ratio proves significant, credit risk factors can be applied selectively to specific groups of assets with particular characteristics in time-series tests. The ultimate goal is to develop a model that, instead of incorporating factors in a binary manner, considers their presence for certain assets based on threshold conditions. This approach aims to maximize explanatory power while adhering to the parsimony principle, ultimately improving the effectiveness of asset pricing models by using factors only where they are most relevant.
 
To evaluate the performance of threshold asset pricing models, data from companies in the Iranian capital market (2001-2023) were used after applying common filters. Threshold regression techniques investigated the threshold effect, and the GRS test along with A(|aᵢ|), A(|aᵢ|/|r̄ᵢ|), and AR², tested the research hypotheses.
 
The results demonstrate that the distance-to-default variable, both individually and in conjunction with other characteristics, is influenced by the threshold effect of the debt ratio. Specifically, there is a significant negative relationship—where a lower distance-to-default corresponds to higher expected returns—for stocks with high debt ratios. This indicates that for companies with higher leverage, investors demand greater returns due to the increased risk of default. Utilizing the threshold effect of variables affecting asset returns in cross-sectional regressions allows for a more precise and detailed analysis of how these variables impact individual stock returns. Additionally, incorporating the default risk factor into the examined asset pricing models enhances their explanatory power and predictive ability, especially for test assets with high debt thresholds. This improvement highlights the effectiveness of considering threshold effects in the models, as it increases their ability to accurately predict returns for assets that are more susceptible to default risk due to higher levels of indebtedness.
 
The results suggest that to adhere to the parsimony principle, threshold asset pricing models can be used to develop pricing models and price a subset of test assets with specific characteristics, yielding appropriate results.

In addition to its impact on public health, the COVID-19 pandemic posed significant challenges to global economies, particularly stock markets. Consequently, the pandemic's effects on stock markets opened diverse avenues for research. Scholars worldwide have sought to explore this topic from various perspectives. This article aims to conduct a bibliometric analysis of studies on COVID-19 and the stock market in Iran by mapping and analyzing scientific research trends in this domain.

This applied study employs a bibliometric approach. The statistical population comprises all articles published on the subject of COVID-19 and the stock market in Iran between February 2020 and January 2023. Data analysis was conducted using Excel, Ravar PreMap, and VOSviewer software.

Considering the process of publishing articles in the field of COVID-19 research and the stock market in Iran, the results show that the trend of publishing articles in this field is upward, with the highest number of published articles in 2022, accounting for 26 titles and 50% of all articles. Regarding the intellectual structure of research in the field of COVID-19 and the stock market in the country, the evaluation of the intellectual structure in bibliometric analysis is usually done by examining the most frequent keywords and their co-occurrence. The most frequent keywords help identify the main topics that make up the intellectual structure of the research and word co-occurrence is used to identify clusters, track time trends, and analyze the density of keywords. These methods help introduce dominant, developed, saturated, extinct, or emerging topics. In this research, to investigate the most frequent keywords, the main words used in the title, abstract, and the keywords selected by the authors in each article were analyzed. In total, researchers in the field of COVID-19 and the stock market in Iran have used 312 main keywords in their articles. The keywords "COVID-19" (47 occurrences) and "stock market" (42 occurrences) are the most frequent in this field, followed by "financial markets," "pandemic," "crisis," "stock market index," "stock market return," and "stock return." These keywords form the main core of the topics that make up the intellectual structure of research in this field in Iran. To investigate word co-occurrence, social network mapping was used. Analysis of the social network map of word co-occurrence was conducted using three methods 1) Analysis of clusters of the most frequent keywords, 2) Investigation of the time process of the formation of clusters of the most frequent keywords, and 3) Density analysis of the most frequent keywords. However, due to the newness of the field and the small number of articles in this area in Iran, analyzing the articles by examining the time trend and the density of the most frequent keywords would not be meaningful. Therefore, word co-occurrence was analyzed only through the examination of clusters of the most frequent keywords. Drawing the word co-occurrence map in the field of COVID-19 research and the stock market in Iran identified only one thematic cluster. This cluster includes eight main keywords: COVID-19, stock market, financial markets, epidemic, crisis, stock market index, stock market return, and stock return, all connected by red lines. Thus, it can be said that the keywords in this cluster indicate the direction of studies in the field of COVID-19 and the stock market in Iran, which has developed in recent years. Regarding cooperation networks among universities in the field of COVID-19 research and the stock market in Iran, the results show that although there has been good cooperation between universities in this field, more cooperation is expected, especially with universities in other countries.

Facilitating the process of publishing articles in this field in Iran, it is expected that the results and suggestions of this research will serve as a roadmap for researchers to explore potential research opportunities in this field.

In recent years, there has been a paradigm shift from traditional approaches to behavioral finance perspectives. Behavioral finance, or colloquially investor sentiment, has gained popularity as a modern approach to asset pricing due to empirical support for the topic. There has been a renewed interest in understanding investor behavior in emerging stock markets. Accordingly, the present study aims to identify the asymmetric effect of investor sentiment on trading value and stock returns in the Tehran Stock Exchange.

This study's statistical population includes all companies listed on the Tehran Stock Exchange from 2016 to 2022. The number of these companies varied on different days during the period due to initial public offerings, transfers between the OTC and the stock exchange, and delistings. The Arms index also called the short-term trading Index, was used to measure investor sentiment, and the exponential GARCH model was used to measure sentiment fluctuations. Also, the transaction value and stock return models were estimated using mixed data-based methods.

According to the findings, investor sentiment had a positive effect on transaction value. The effect of investor sentiment was greater when sentiment was positive than when it was negative. Also, sentiment fluctuations had a positive and significant effect on transaction value. On the other hand, the asymmetric effect of investor sentiment fluctuations on transaction value was significant. The results indicated that the effect of positive sentiment fluctuations was greater than the effect of negative sentiment fluctuations on transaction value. On the other hand, the increase in the dollar exchange rate, as a control variable, had a direct effect on transaction value, leading to its increase. Also, the effect of growth in fixed-income fund trading, as a control variable, was negative and significant. According to the results of estimating the stock return equation, with increasing pessimism and rising negative sentiment in the market, stock returns decrease during the period under review and vice versa. On the other hand, the dummy variable coefficient for investor sentiment is significant, which indicates the asymmetric effect of sentiment fluctuations on stock returns. The findings suggest that the effect of positive sentiment on stock returns is greater than that of negative sentiment. In addition, investor sentiment fluctuations have a positive and significant effect on stock returns. The asymmetric effect of investor sentiment risk on transaction value was significant. Finally, when investor sentiment was negative, the effect of sentiment fluctuations on stock returns was greater than when investor sentiment was positive in the market.

According to the results, the following hypotheses were not confirmed: the effect of positive sentiment fluctuations on transaction value is greater than that of negative sentiment fluctuations, and when investor sentiment is negative, the effect of sentiment fluctuations on stock returns is greater than when investor sentiment is positive in the market. However, the other research hypotheses were confirmed.