Abstract

This study investigates the relationship between consumer confidence and firm-level stock price volatility in Colombia, focusing on 2015–2025. Stock market volatility, often heightened in emerging economies due to structural weaknesses such as low liquidity and investor concentration, poses significant risks to financial stability. Consumer confidence, a behavioral indicator reflecting household perceptions of economic conditions, is increasingly recognized as a market sentiment and investment dynamics determinant. Despite its importance, limited empirical evidence exists on its influence within the Colombian equity market. This research addresses this gap by applying panel data regression techniques to a dataset of twenty publicly listed firms across diverse industries.

The analysis incorporates quarterly data on stock returns, consumer confidence, firm size, leverage, and policy rates, with interaction terms included to test moderation effects. Results reveal a significant negative relationship between consumer confidence and stock volatility, indicating that higher confidence stabilizes equity returns. Policy rates, conversely, exhibit a strong positive effect, underscoring the role of monetary policy in amplifying volatility. Firm size and leverage show no significant moderating effects, suggesting that the impact of consumer confidence is broadly consistent across firms, irrespective of structural characteristics. Diagnostic and robustness tests, including stationarity checks, heteroscedasticity corrections, and alternative specifications, confirm the stability and reliability of these findings.

The study contributes to the literature on behavioral finance in emerging markets by demonstrating the stabilizing role of consumer confidence at the firm level, challenging conventional efficient market assumptions. The results provide insights for investors, who can use confidence indicators as predictive tools for risk management, and for policymakers, who must consider the volatility implications of monetary interventions. The findings highlight the dominance of aggregate behavioral and macroeconomic factors over firm-specific attributes in shaping volatility within Colombia’s equity market.

Abstract i

Dedication ii

Acknowledgements iii

INTRODUCTION 1

1.1 Background to the Study 1

1.2 Statement of the Problem 2

1.3 Research Objectives 3

1.3.1 General Objective of the Study 3

1.3.2 Specific Objectives of the Study 3

1.4 Research Questions 3

1.5 Justification of the Study 3

1.6 Significance of the Study 5

1.7 Scope of the Study 6

1.8 Limitations of the Study 7

LITERATURE REVIEW 8

2.1 Introduction 8

2.2 Theoretical Framework 8

2.3 Stock Market Volatility in Emerging Markets 9

2.4 Consumer Confidence and Market Dynamics 10

2.5 Firm-Specific Moderators: Leverage and Size 11

2.6 Panel Data Empirical Literature 12

2.7 Research Gap 13

2.8 Conceptual Framework 14

METHODOLOGY 15

3.1 Introduction 15

3.2 Research Design 15

3.3 Data Sources and Description 17

3.4 Variables and Measurements 19

3.5 Model Specification 20

3.6 Estimation Techniques 22

3.8 Ethical Considerations 25

RESULTS AND ANALYSIS 27

4.1 Introduction 27

4.2 Descriptive Statistics 27

4.2.1 Volatility 27

4.2.2 Colombia Consumer Confidence Index 29

4.2.3 Firm Size (Moderating Variable) 30

4.2.4 Leverage 32

4.2.5 Policy Rate 34

4.3 Correlation analysis 35

4.4 Diagnostic Tests 36

4.4.1 Stationarity Tests 36

4.4.2 Multicollinearity test (VIF) 37

4.4.3 Heteroscedasticity test (Wald) 38

4.4.4 Autocorrelation test (Wooldridge) 38

4.4.5 Cross-sectional dependence test (Pesaran CD) 39

4.4.6 Specification test (RESET) 40

4.4.7 Remedies applied 40

4.5 Regression Analysis 42

4.5.1 Baseline Model (CCI only + control) 42

4.5.2 Extended Model 43

4.5.3 Interaction 44

4.6 Robustness Checks 45

4.6.1 Alternative Measures of Volatility, Firm Size, and Leverage 45

4.6.2 Sub-sample Analysis (Different Time Periods) 46

4.6.3 Outlier and Influential Observation Checks 47

4.6.4 Stability of Results 47

4.7 Discussion 48

4.8 Chapter Summary 51

SUMMARY, RECOMMENDATIONS, AND CONCLUSION 53

5.1 Summary 53

5.2 Policy and Practical Implications 53

5.3 Recommendations for Future Research 54

5.4 Conclusion 55

References 56

INTRODUCTION

1.1 Background to the Study

Stock market volatility is defined as the extent of fluctuations in stock prices, and over time, it is identified as one of the leading indicators of market risk. In emerging markets, volatility is generally higher than in the more developed markets as the markets are smaller, have fewer investors, and are more open to world economic shocks (Tabash et al., 2024). This increased volatility can influence investment, financial stability, and economic growth. It is thus essential to determine the source of the volatility in the emerging economies to enhance market resistance and policy interventions.

Consumer confidence measures the household perception of present and future economic conditions. It is an essential indicator of possible transitions in consumption and investment patterns. In the emerging markets, the changes in consumer confidence can affect the flow of funds into the financial markets, leading to fluctuations in stock prices (Chikwira & Mohammed, 2023). Gaining insights into how consumer confidence interacts with the stock markets is vital to investors and policymakers who want to predict and manage high-volatility periods.

In Colombia, the stock market is a constituent of the financial system, contributing to economic growth by mobilising savings and helping create capital. The main equity index, COLCAP, tracks the performance of 20 companies in Colombia's equity market (Garay & Pulga, 2021). However, the market remains relatively small and illiquid, compared to those in developed economies, and thus is vulnerable to sudden changes in economic indicators. Despite this, few empirical studies have been performed on the domestic factors affecting Colombia's stock market volatility, such as consumer confidence.

1.2 Statement of the Problem

Volatility in the stock market is a critical concern to investors, policymakers, and regulators due to its implications for financial stability and economic growth. In developing markets like Colombia, volatility tends to be higher due to structural factors such as low market liquidity, small stock market investors, and higher vulnerability to external shocks (Yang et al., 2021). Gaining insight into what causes volatility in such markets is necessary to formulate risk management strategies and facilitate the effective operations of financial systems.

Consumer confidence indicates how households view the economic situation and their future expectations. Consumer confidence in many countries is connected to fluctuating economic activity, such as behavioural consumption, investments, and savings (Koskelainen et al., 2023). These changes can also affect the financial markets, causing investment and stock price variations. However, in Colombia, empirical studies analysing the impact of consumer confidence on stock market volatility, especially at the firm level, are limited. The existing studies in Latin America have concentrated on the broad macro-indicators or general market index measures without evaluating firm-specific details, which can contribute to an increase or reduction in volatility.

The absence of evidence leads to a gap in explaining the relationship between consumer confidence and the volatility of the stock price in Colombia. This knowledge is essential because it can guide investors interested in managing portfolio risks and policymakers interested in developing policies to increase market resilience (Sutton, 2025). Addressing this gap, the current study explores how consumer confidence affects the volatility of individual company stocks in Colombia's equity market from 2015 to 2025. It also assesses how the size and leverage as firm-level factors might moderate this relationship, offering a more refined view of volatility patterns in the Colombian equity market.

1.3 Research Objectives

1.3.1 General Objective of the Study

The general objective of this study is to evaluate the influence of consumer confidence on the stock market in Colombia.

1.3.2 Specific Objectives of the Study

1. To examine the relationship between consumer confidence and individual company stock volatility.
2. To assess the effect of the size of firms on the relationship between consumer confidence and stock volatility.
3. To analyse the moderating effect of leverage on the relationship between consumer confidence and stock volatility.
4. To provide empirical evidence on the role of domestic economic indicators in explaining Colombia's stock market volatility from 2015 to 2025.

1.4 Research Questions

1. What is the relationship between consumer confidence and individual company stock volatility?
2. What is the effect of the size of firms on the relationship between consumer confidence and stock volatility?
3. What is the moderating effect of leverage on the relationship between consumer confidence and stock volatility?
4. What empirical evidence shows the role of domestic economic indicators in explaining Colombia's stock market volatility from 2015 to 2025?

1.5 Justification of the Study

The stock market's volatility is a significant issue in financial markets due to its impact on investment practices, portfolio management, and the general stability of the economy. In developing economies like Colombia, volatility is often higher because of structural issues like less market liquidity, small bases of investors, and greater exposure to domestic and external shocks (Yang et al., 2021). It is essential to identify the factors that lead to the volatility of such markets to enhance the financial system's resilience and create an environment that facilitates long-term economic growth. This study examines consumer confidence as a potential cause of stock price volatility in the Colombian market from 2015 to 2025.

Even though consumer confidence is extensively considered a valuable indicator of household economic expectations, few empirical studies have been done to assess its impact on stock market volatility in Colombia. Most of the research-based literature in the Latin American setting has focused more on the macroeconomic factors or the analysis of volatility trends at the market index level, without evaluating variations on the firm level (Akgüller et al., 2025). This generates a knowledge gap in understanding the relationship between domestic economic indicators like consumer confidence and the risk characteristics of individual companies' stocks. Exploring these dynamics can provide more details on volatility patterns and offer evidence that market participants could use to make informed decisions.

This study is significant since it aims to contribute practically and academically. In the case of the investor, it offers helpful information on how consumer confidence may impact the volatility of individual stocks, enabling them to manage risks in their portfolio more effectively (Léber & Egyed, 2025). For policymakers and regulators, the study identifies domestic economic factors that may influence the fluctuation of stock prices, hence the implementation of stabilising initiatives in the financial markets. Academically, the research adds value by enhancing current knowledge regarding research on the firm-level volatility in the emerging market, which has not been heavily researched, especially in the context of the Colombian equity market.

1.6 Significance of the Study

This study is significant since it examines the impact of consumer confidence on the volatility of stock markets in Colombia, an emerging economy whose financial markets are more prone to domestic and external shocks. It is essential to know about this relationship so that investors can make informed decisions in a market highly characterised by frequent price fluctuations. By exploring the stocks of 20 individual companies instead of the market indices, this study offers a clear insight into firm-level volatility patterns, providing practical insights for risk management of investment performance in the context of individual companies. According to Akin & Akin (2024), recognising the effect of consumer confidence on volatility also allows authorities to forecast when the risk will be very high and develop interventions that reduce the impact of extreme pricing variations. The knowledge gathered from this study can be applied to developing fiscal and monetary policies that will boost investor confidence and guarantee a more secure financial system.

Academically, the study contributes to the limited literature on volatility at the firm level for emerging economies in Latin America. Empirical studies in the region have focused either on macroeconomic variables or market-level indicators, without regard for how factors at the firm level influence domestic measures like consumer confidence (Bitetto et al., 2023). By making this contribution, the study sets the stage for further empirical studies of drivers of volatility in other emerging markets. This renders the study relevant to Colombia and comparative studies in economies with the same characteristics.

1.7 Scope of the Study

The study focuses on the relationship between stock market volatility and consumer confidence in Colombia. It uses quarterly data from 2015 to 2025 from 20 companies in Colombia's equity market. The research scope includes the calculation of firm-level volatility from stock returns and measuring consumer confidence from the Colombian Consumer Confidence Index. With a focus on the Colombian equities market, the research hopes to help improve knowledge regarding how domestic economic factors affect the movements in stock prices in an emerging market context, where financial markets are less developed and more volatile.

The research utilises firm-specific characteristics such as size and leverage as moderator variables to explore how these characteristics influence the interaction between consumer confidence and stock volatility. Macro control variables like real interest rates and credit default swap spreads are also utilised to account for economy-wide impacts. The inclusions ensure the analysis provides micro- and macro-level impacts in the Colombian stock market. This scope allows for an extensive and multi-dimensional examination of drivers of volatility in the country's financial market, complementing gaps in previous research that focused mainly on indices of markets alone.

The study is confined to secondary data that uses credible sources like Refinitiv and Banco de la República. It excludes companies whose data are extensively missing within the period used to conduct the study to uphold the robustness of the panel dataset. Although the results are restricted to Colombia, they might not be directly applied in other markets with different structural and economic characteristics. Future research may focus on comparative research of many emerging markets or other firm-level factors like industry classification and corporate governance indicators to gain a broader perspective on the elements that influence stock market volatility.

1.8 Limitations of the Study

This research only concentrates on the Colombian stock market, and its findings may not directly apply to other emerging or developed markets. Various countries' structural and economic aspects imply that the association between consumer confidence and stock market volatility may differ in various situations (Ghosh, 2022). The results should therefore be interpreted in the context of Colombia's specific characteristics of the equity market. In the future, it is possible to expand the analysis by comparing more than two markets to identify similar trends in the economies with various levels of financial development, investor participation, and market stability.

The study uses secondary data from established databases such as Refinitiv and the Central Bank of Colombia. Even though these sources are considered reliable, there can be a problem of accuracy, consistency, or data availability during the ten years of the research. The analysis does not include firms with extensive missing data to avoid its weakness, which could, to some extent, affect the generalizability of the results (Enders, 2022). Also, the research uses quarterly data to balance the granularity and the number of observations. However, data on a higher frequency scale may present more information about the short-term volatility tendencies.

In this study, an analysis using only quantitative variables is conducted. Qualitative variables like investor behaviour, political events, or external global shocks affecting the stock market's volatility are not integrated. These factors are beyond the scope of the present research and, therefore, may help better understand market fluctuations if examined in future research. Including quantitative and qualitative methods in future studies would add to the analysis and provide better ideas on what underlies market volatility, especially in markets where non-economic factors could be significant in determining investor actions and stock prices (Adamyk et al., 2025).

2.1 Introduction

This literature review provides a theoretical and empirical foundation for examining the degree to which consumer confidence influences firm-level stock price volatility. It begins by outlining the major theory employed in explaining consumer confidence, primarily behavioural finance, and contrasting it with orthodox models such as the Efficient Market Hypothesis. Subsequently, it describes the nature of volatility in emerging markets and the specific characteristics that render them susceptible to consumer-based fluctuations. Subsequent sections review the contribution of consumer confidence towards market dynamics from both developed and emerging market literature. The review also considers firm-level variables like size and leverage, which can moderate the relationship between consumer confidence and volatility. The chapter concludes by assessing the appropriateness of panel data methodologies for capturing these relationships across firms and over time. Through a combination of behavioural, structural, and econometric viewpoints, this chapter highlights significant literature gaps and rationalises the necessity of a firm-level panel study of Colombia's equity market from 2015 to 2025.

2.2 Theoretical Framework

This study is theoretically founded on behavioural finance that challenges rational behaviour and the efficient market that prevails in conventional financial models. Behavioural finance identifies that psychological and emotional influences play a role in investment decisions and contribute to mispricing and volatility (Akin & Akin, 2024). In this regard, consumer confidence is used as a proxy for investment. As confidence levels are high, investors may get too optimistic, leading to overvaluation and risk underestimation. On the other hand, poor consumer confidence may create fear or risk aversion, with volatility being exaggerated. According to Akin and Akin (2024), these variables could impact market behaviours. This opposes the Efficient Market Hypothesis (EMH), which presumes that the prices capture all data and volatility is the only outcome of new, unexpected information (Mikołajek-Gocejna & Urbaś, 2023).

Behavioural finance is relevant in emerging markets like Colombia, where companies operate under incomplete information, reduced investment choice, and an underdeveloped financial infrastructure. These conditions maximise the influence of consumer-based behaviour, a key area to study. While EMH assumes that irrationalities of individuals cancel out in the market, behavioural theory suggests that persistent changes, such as shifts in consumer confidence, can create consistent mispricing (Mikołajek-Gocejna & Urbaś, 2023). This study uses the behavioural approach and concentrates on investigating the role of consumer confidence on firm-level volatility, considering that individual stocks can respond differently based on firm characteristics. The study also utilises behavioural explanations as justification for using a panel data methodology, which allows for the analysis of cross-sectional and time-series variation in price history (Rehan et al., 2023). Thus, the behavioural finance framework provides a proper insight for examining volatility patterns in the Colombian equity market.

2.3 Stock Market Volatility in Emerging Markets

Stock market volatility refers to the extent to which stock prices fluctuate with time, typically due to uncertainty or risk in financial markets. In developed economies, volatility is generally relatively repressed because of more profound markets, advanced financial instruments, and more knowledgeable investor bases. However, developing economies like Colombia undergo increased volatility due to low participation by investors, low liquidity, weaker institutions, and higher exposure to international events (Hoang & Mateus, 2024). Such structural characteristics enhance the impact of economic shocks and consumer confidence and make volatility a more intricate and long-lasting problem. The Colombian stock market, while developing, is relatively illiquid and small and thus more susceptible to local and international shocks (Bonilla-Mejía & Villamizar-Villegas, 2022). These characteristics justify the requirement of research at the local level, identifying the determinants of stock volatility in such weak markets.

Existing literature examining volatility in emerging markets has focused on macroeconomic variables or market indices in aggregate terms using GARCH-type models to capture volatility. While informative, such approaches overlook the role of behavioural variables and heterogeneity at the firm level. Index-level analysis also obscures how firms respond individually to economic shocks and investor behaviour (Kijkarncharoensin, 2025). Therefore, the firm-specific volatility dynamics are poorly understood, especially in the Latin American market. This creates a gap. Very few studies have considered the role of micro-level behavioural or structural variables, such as consumer confidence, in the individual stock's volatility. This research seeks to bridge that gap by using firm-level data to examine how consumer confidence, as a behavioural factor, influences stock volatility of Colombian firms, thus offering a more detailed description of market dynamics in an emerging economy.

2.4 Consumer Confidence and Market Dynamics

Consumer confidence is the general optimism or pessimism conveyed by households about the current and future state of the economy. It is becoming more widely used in finance to determine investor behaviour. Changes in consumer confidence can lead investors to change their expectations for return and risk, affecting trading behaviour or the price index (Padmavathy, 2024). Various studies have found that higher consumer confidence is associated with higher stock prices and lower volatility in the developed markets. In contrast, declining confidence is associated with market corrections or increased volatility (Akin & Akin, 2024). These effects are driven by emotional responses rather than new economic data, which aligns with the behavioural finance view. When markets are made of price movements and variations, they cause uncertainty, resulting in disproportionality.

Consumer confidence's influence on market forces in emerging markets is not well known due to data limitations, and the existing studies are focused less on this. Research by Chikwira and Mohammed (2023) examined the market performance concerning macroeconomic changes but omitted variables like confidence indices. Other studies have centred on whether interest rates, inflation, or external shocks cause volatility in Colombia and other Latin American markets. Not many researchers included consumer confidence, though it is an important driver of investor decisions. Because both external and internal shocks apply to Colombia, investigating how consumer confidence influences stock market performance can offer important insights (Akin & Akin, 2024). This study bridges the gap by empirically examining the role of consumer confidence in firm-specific volatility to identify how psychological forces determine price movements in an emerging economy with structural and behavioural constraints.

2.5 Firm-Specific Moderators: Leverage and Size

While macroeconomic factors influence market events, firm-specific traits mediate their effects. In this regard, two powerful moderators are firm size and leverage. Larger companies are less susceptible to global shocks because they probably have more stable investor bases, better corporate governance systems, and more diversified revenues. Limited resources, concentrated ownership structures, low market liquidity, and volatility are characteristics of small businesses. Bitetto et al. (2023) contend in their study that investors' overreactions and global economic shocks are more likely to affect small businesses. Smaller companies are more susceptible to a big impact from a price change, and the size of the business can also influence how confident customers are about volatility.

A key moderating factor in market volatility is leverage in the company's capital structure. Companies with high leverage levels are also more vulnerable to changes in funding risk and policy rates. Therefore, they are more prone to dumping their shares in panic when consumer confidence is low. Investors can view high leverage as a sign of weakness, and concerns are being raised about an aggressive portfolio. Fu et al. (2021) found that more leveraged companies were disproportionately affected when financial distress was more severe. By including the CCI x Firm Size and CCI x Leverage interaction terms in the model, this study seeks to provide the answer to this question: under what structural conditions do the CCI x Firm Size and CCI x Leverage interaction terms result in the association between consumer confidence and firm-specific stock volatility? This provides a better understanding of the behavior dynamics at the firm level that can be applied to Colombia's policymakers, researchers, and market players.

2.6 Panel Data Empirical Literature

Panel data analysis has some unique strengths in financial research, particularly when it is employed to study the behaviour of firms over time. Relative to cross-sectional or time-series methods, panel data models can control for unobserved heterogeneity, such as firm-specific attributes that change over time but across different firms. This is especially useful in analysing how an aggregate factor, such as consumer confidence, affects firms differently depending on their attributes. Fixed effects models, for example, allow the distinction between within-firm effects by controlling for time-invariant attributes. Random effects models can also account for firm-level variation under both observed and unobserved factors, improving estimation efficiency.

The panel data methods have been utilised in emerging economies to investigate problems like investment conduct, risk-taking behaviour, and financial development. For example, research by Ranaweera and Kawshala (2022) used panel regressions to test the influence of financial knowledge on investments in the Colombo Stock Exchange. However, few Latin American studies have applied panel techniques in examining behavioural finance questions. They employ primarily single-index regressions or aggregate time-series specifications that limit the ability to detect heterogeneity in response patterns. Methodologically, this study contributes to the research by making a panel of 20 Colombian firms for 2015-2025 to examine firm-level volatility based on consumer confidence and structural factors. With the application of panel data, the study provides a more detailed and statistically more robust analysis of the relationship between stock market vitality and consumer confidence in the Colombian equity market.

2.7 Research Gap

A significant research gap from reviewing the existing literature is a critical insight into how behavioural indicators like consumer confidence influence emerging markets' stock volatility at the firm level. Most studies employ market indices that aggregate performance and suppress firm-specific responses to price shifts. Most Latin American research excludes psychological considerations of investment behaviour and concentrates on traditional macroeconomic determinants such as inflation, policy rates, and exchange rate volatility (Movsisyan et al., 2024). While these are undoubtedly important, they cannot explain how emotional and cognitive biases influence investor preference and generate price movements. Failure to incorporate behavioural analysis in volatility models diminishes the power to define such models in emerging markets.

Even for those studies that consider sentiment indicators, there is little application of interaction effects to test whether firm characteristics impact how consumer confidence impacts volatility. Variables like firm size and leverage are all commonly known to be important in finance literature, but their roles as moderators of behaviour effects are poorly researched in Latin America (Yu, 2023). There is also a methodological shortcoming, as most of the research in this area employs single-equation time-series models rather than panel data, which can account for both firm-specific and temporal variation. This study fills these gaps by applying behavioural theory and empirical panel data techniques to examine the impact of consumer confidence on firm-specific volatility in Colombia. It also adds to this by exploring interaction effects and providing a more advanced description of risk behaviour in an emerging market context.

2.8 Conceptual Framework

The study's conceptual framework is based on empirical research on the variables influencing behavioral finance and volatility. As an indicator variable, it assumes that consumer confidence directly impacts the volatility of firm-level stock returns. The fundamental premise is that investor behavior shifts in response to shifts in consumer confidence and is probably related to impacts on market stability. These responses vary from firm to firm and are related to the structural features of the firm, including size and leverage. Given that they may be more susceptible to price fluctuations than small businesses, some large companies are thought to be more stable and less likely to experience investor overreactions. Similarly, the more leveraged a company is, the greater its financial risks, and fluctuations in investor behavior may lead to high price volatility.

Policy rates are a control variable that accounts for macroeconomic factors that could undermine consumer confidence, and the correlation of stock volatility in a specific company that makes up the framework. In order to examine how the impact of consumer confidence varies with firms that are differentiated based on their profiles, the empirical model incorporates the panel regression model with the interactions, CCI x Size and CCI x Leverage. Through this procedure, the framework enables the examination of both the conditioned influence and the direct impact of consumer confidence on cost motion. The strategy improves the explanatory power of the model and supports the theoretical hypothesis that the behaviour of investors varies. The conceptual framework presents a structured model to study the behavioural-financial relations in the equity market in Colombia, from 2015 to 2025.

METHODOLOGY

3.1 Introduction

The chapter introduces the methodology for analyzing Colombia's relationship between consumer confidence and stock market volatility. The quantitative approach applies panel data regression techniques to capture time-series and cross-sectional variation differences in selected firms in the Colombian equity market from 2015 to 2025. The design enables a broader analysis of volatility drivers by accounting for the differences between firms and transitions over time. The chapter starts by explaining the research design and why panel data is the best approach for this study. It then outlines the data sources, the selection of companies, and the construction of the dataset, including calculating volatility and measuring consumer confidence. Variables, which include control and moderating factors, are clearly stated, followed by model specification and reasons for including interaction terms. It also discusses the estimation techniques, diagnostic tests, and robustness tests employed to determine statistical validity. Ethical considerations regarding the secondary data usage are also discussed to maintain transparency, reliability, and academic integrity in the research process.

3.2 Research Design

This study uses a quantitative research design to examine the correlation between consumer confidence and the volatility of stock markets in Colombia. Quantitative methods are suitable for testing statistical relationships between measurable variables using numerical data. The study utilizes a panel data model, which combines time and cross-sectional aspects to study several companies simultaneously over multiple periods. This method is selected since it is more accurate in estimation, has more degrees of freedom than that based on purely time-series analysis or only on cross-sectional, and accounts for the unobserved heterogeneity across firms. Using quarterly data from 2015 to 2025, the study has sufficient data to determine the short-term and medium-term volatility patterns.

Panel data analysis is particularly significant in financial market research because it can observe each firm's characteristics when assessing common influences like consumer confidence. The study investigates twenty Colombian Stock Exchange firms selected based on the availability of data and persistent trading volume during the study duration. Firm-level volatility is computed based on quarterly log return standard deviation for the dynamic measurement of price movements. This method offers a more realistic reflection of volatility than yearly figures, which can reflect distorted extreme variations.

The dependent variable in the analysis is firm-level stock volatility, and the key independent variable is the Colombia Consumer Confidence Index (CCI), which is extracted from Refinitiv. Firm size and leverage are also included as the moderating variables to test whether firm features influence the effect of consumer confidence on volatility. Policy rates are also included as control variables to account for the macroeconomic factors. The research design involves interaction terms (CCI × Size and CCI × Leverage) for testing conditional effects to provide a more nuanced understanding of volatility determinants.

The model specification is underpinned by a panel regression framework such that, depending on the result of the Hausman test, fixed or random effects estimation can be employed. Diagnostic procedures encompassing heteroskedasticity tests, serial correlation, and multicollinearity tests ensure the model's validity. Robustness tests, such as alternative volatility measures and sub-period tests, are conducted to ensure the stability of the findings. The systematic design ensures that the study is aligned with the objectives of discovering, quantifying, and explaining the relationship between consumer confidence and firm-level volatility in Colombia. The systematic nature of the approach ensures the findings are more reliable and applicable in academia and practice.

3.3 Data Sources and Description

This study relies on secondary data from reputable sources for reliability and consistency. Stock market data and financial data for companies, including price history, are obtained from Refinitiv. The data covers ten years from 2015 to 2025 to allow for the inclusion of various economic cycles and varying market conditions. The selection of this period provides sufficient observations for utilization in robust statistical analysis. Consumer confidence data is also sourced from Refinitiv, specifically the Colombia Consumer Confidence Index (CCI). Policy rates are macroeconomic variables sourced from Banco de la República in Refinitiv.

The sample consists of twenty companies listed on the Bolsa de Valores de Colombia (BVC), chosen based on three criteria: long-term listing during the period under study, active trading with adequate liquidity, and complete availability of quarterly data. The selection ensures that volatility measures reflect actual market activity, not illiquidity distortions or data gaps.

The companies include:

1. Acerias Paz del Rio SA
2. Banco Davivienda SA
3. Almacenes Exito SA
4. Bmc Bolsa Mercantil De Colombia SA
5. Bolsa de Valores de Colombia SA
6. Celsia Colombia SA ESP
7. Cementos Argos SA
8. Corporacion Financiera Colombiana SA
9. Ecopetrol S.A
10. Empresa de Telecomunicaciones de Bogota SA Esp
11. Enka de Colombia SA
12. Grupo Aval Acciones y Valores S.A
13. Grupo Bolivar SA
14. Grupo Cibest SA
15. Grupo de Inversiones Suramericana SA
16. Grupo Nutresa SA
17. Interconnection Electric SA ESP
18. Mineros SA
19. Organizacion Terpel SA
20. Promigas SA ESP

The industries that these firms are engaged in are banking, energy, utilities, manufacturing, telecommunications, retail, food processing, aviation, mining, and textiles. The resultant wide-sector coverage is necessary to have diversified industry volatile trends in the analysis. Diversification across industries eliminates shock bias at the sector level, such as oil price fluctuations affecting any energy company, or regulatory changes affecting banks. Moreover, combining large-cap with mid-cap companies allows for determining whether the association between consumer confidence and volatility is affected by the size of firms. Larger companies are also more stable in market changes than smaller ones, but they can be more vulnerable to these changes, making the stock price fluctuate more.

The volatility at the firm level is measured using values obtained by adjusting Refinitiv closing prices to reflect corporate actions, including dividends and stock splits. Such adjustments are essential to avoid distortions in the calculation of returns, such that any market price movements observed correspond to market forces accurately. The quarterly log returns are calculated in each firm, and volatility is computed as the standard deviation of returns in each firm. This method eliminates short-term noise, but does not exclude the technique's sensitivity to sharp market fluctuations. The frequency of the quarterly basis aligns with the consumer confidence data; therefore, there is temporal consistency.

The data provides a multidimensional view of the Colombian equity market by incorporating firm-level market data, macroeconomic indicators, and sector representation. This procedural strategy leads to a balanced panel structure, ensuring all twenty firms have the complete information for each quarter within the study period. This data type best fits the panel regression model because it is efficient and reduces the biases of unbalanced panels. The resultant data becomes the empirical foundation for conducting the tests of the hypotheses applied in this research. It enables a subtle measure of the interconnection between consumer confidence and firm-specific and macroeconomic variables in influencing stock market volatility in Colombia.

3.4 Variables and Measurements

The dependent variable in this study is firm-level stock volatility, which is the standard deviation of quarterly log returns for each firm. Dividends, stock splits, and other corporate actions are taken into consideration by Refinitiv adjusted closing prices. Quarterly log returns are calculated using a natural log of the ratio of consecutive quarter-end prices. The rolling window technique smooths out random short-run volatility but is vulnerable to wide fluctuations in market conditions. The measure captures investors' uncertainty regarding a firm's share performance under different economic conditions.

The key independent variable is the Colombia Consumer Confidence Index (CCI), which is obtained from Refinitiv. The CCI is a quarterly survey indicator that reflects current opinions regarding economic conditions and household future expectations. It measures opinion on employment, personal finances, and the economy in general. The index is a percentage, where positive values illustrate optimism and negative values demonstrate pessimism. The quarterly frequency is applied to measure volatility directly. This study uses the raw CCI score without transformation to preserve interpretability and explore directly the effect of changes in consumer confidence on stock volatility at the firm level.

There are two moderator variables: firm size and firm leverage. Firm size is measured as the natural log of market capitalization, calculated from the number of shares outstanding multiplied by market price per share. This transformation reduces skewness and makes comparing companies of different sizes possible. Company leverage is calculated as the debt-to-equity ratio from Refinitiv's balance sheets, dated every quarter. These variables test whether the effect of consumer confidence on volatility differs by company characteristics, with larger companies being more stable and highly levered firms possibly facing higher risk exposure.

Control variables are policy rates, which reflect macroeconomic conditions. Policy rates are grouped with interest rate, daily, Banco de la República, daily, Colombia (acocbr), using Banco de la República data in Refinitiv. It is added to control for the effects of external factors on volatility beyond consumer confidence, making it a better estimation of the main relations of interest in the model.

All the variables are measured quarterly from 2015 to 2025 to ensure a balanced and consistent panel dataset for all twenty firms. This synchronization eliminates time-mismatch errors that may distort findings. The data set integrates firm-level and macroeconomic variables into a unified framework for robust panel regression analysis. With the proper definition and measurement of every variable, the study ensures replicable and understandable outcomes. The systemic design for measurement allows for empirical hypothesis testing for the association between consumer confidence, firm characteristics, and stock market volatility in Colombia.

3.5 Model Specification

This study adopts a panel data regression framework to examine the effect of consumer confidence on stock volatility at the firm level, with firm size and leverage as moderating variables. Panel data is particularly suitable for this study because it can capture cross-sectional and time-series differences, accounting for the control of unobserved time-invariant firm-specific characteristics. This dual nature enhances estimation efficiency and provides a richer dataset than an entirely cross-sectional or time-series approach. A balanced panel of twenty listed public companies in Bolsa de Valores de Colombia (BVC) from 2015 to 2025 consistently observes each company throughout the study. The baseline model concentrates on the direct relationship between consumer confidence and volatility and accounts for the firm-level variables such as size, leverage, and macroeconomic variables such as policy rates. By including these variables, the model aims to isolate the unique influence of consumer confidence on volatility in the Colombian equity market.

The baseline model is expressed as:

VOL_it​=β₀​ + β₁​CCI_t ​+ β₂​SIZEit​ + β₃​LEV_it​ + β₄​PR_t​ + β₅​CDS_t​ + ϵ_it​

Where:

• VOL_it​​ = Firm-level volatility for firm i at time t
• CCI_t ​​ = Consumer Confidence Index at time t
• SIZE_it​ = Firm size (log of market capitalization) for firm i at time t
• LEV_it​ = Firm leverage (debt-to-equity ratio) for firm i at time t
• PR_t = Real interest rate at time t
• CDS_t​ = Colombian sovereign Credit Default Swap spread at time t
• ϵ_it​ ​ = Error term

Regarding examining moderating effects, the analysis is evaluated beyond the baseline model by inserting the interaction terms between the consumer confidence and the moderating variables (Firm size and leverage). The terms enable the model to determine whether the impact of consumer confidence on volatility alters with the change in the company’s size or debt structure. Larger businesses, for example, can withstand shocks if they have sufficient resources and are well-established. On the other hand, when confidence declines, highly leveraged businesses may experience the effects more acutely. The analysis of the relationships is more thorough when the main and interaction effects are included. The Hausman test is used to assess which specification is preferred to provide greater statistical validity and efficiency of the result after testing both fixed and random effects estimations.

VOL_it​=β₀​ + β₁​CCI_t ​+ β₂​SIZEit​ + β₃​LEV_it​ + β₄​(CCI_t​×SIZE_it​) + β₅​(CCI_t​×LEV_it​) + β₆PR_t​ + β₇​CDS_t + ϵ_it​

3.6 Estimation Techniques

The study makes use of regression analysis with panel data. It investigates the relationship between Colombian firm stock volatility and consumer confidence, with firm size and leverage as moderator variables. Panel data is used because it can analyze time-series variation over time and cross-sectional variation across firms, yielding more accurate and useful estimates. The time-invariant unobserved firm-level influences, such as management, industry specificity, or geographic scope, can be controlled by the model using this approach. Available from 2015 to 2025, a balanced panel data sample of 20 securities listed on Colombia's Bolsa de Valores de Colombia (BVC) ensures that the same companies are tracked throughout all time periods and that there are no excessive missing data points that could skew or impact the estimation of the parameter. The model uses macroeconomic control (policy interest) and microeconomic predictors (firm size and leverage) for greater economic impact. The research thoroughly investigates volatility determinants at the Colombian market's firm level using these analysis levels.

The two primary estimation methods are the Fixed Effects (FE) and Random Effects (RE) models. The FE model captures firm-specific intercepts and hence controls for unobserved, time-invariant characteristics that may be correlated with the explanatory variables. This is preferable in finance market studies where omitted variables lead to biased conclusions. In contrast, the RE model assumes that firm-specific effects are random and not correlated with the regressors, giving more efficient parameter estimates under this assumption. A Hausman test is employed to determine the most fitting model. This statistical test evaluates whether there is a systematic difference in the relative coefficient estimates between the FE and the RE models. The RE assumption has failed significantly, and the FE model is preferred. Formal testing of this by the study confirms that the chosen model specification is statistically sufficient and fitting to the data set.

The estimation process is done in two steps. In the initial step, the baseline model is estimated to evaluate the direct association between consumer confidence and stock volatility while controlling for firm size, leverage, and macroeconomic conditions. This step sets the basic associations without the added complexity of interaction effects. The second step includes interaction terms between the consumer confidence and the moderator variables, so the impact of confidence can be tested to see whether it varies by firm size or leverage. This sequential process makes interpretation easier since it separates direct from conditional effects. Statistical measures like the coefficient of determination (R²), adjusted R², and the F-statistic are used to test joint significance on model performance. The Bayesian Information Criterion (BIC) and the Akaike Information Criterion (AIC) are also used to compare model fit across specifications. This controlled estimation technique ensures the analysis's statistical and theoretical soundness, which offers a solid basis for interpreting the results in later chapters.
3.7 Diagnostic and Robustness Tests

Diagnostic tests ensure the estimated panel regression models meet the assumptions for proper inference. In pre-estimation diagnostic testing, panel unit root tests like the Levin–Lin–Chu (LLC) and Im–Pesaran–Shin (IPS) tests are used to check for variable stationarity. Non-stationary variables can result in spurious relationships, where apparent correlations are due to shared trends rather than substantive linkages. Where non-stationarity is detected, differencing or transformation is applied. Multicollinearity among explanatory variables is tested using the Variance Inflation Factor (VIF), and values over 10 indicate potential problems. Post-estimation diagnostics after estimation test for heteroscedasticity using the modified Wald test and autocorrelation using the Wooldridge test. If applicable, robust or cluster-robust standard errors are utilized to correct standard error estimates. These steps ensure the statistical environment is specified correctly before concluding the model results.

Apart from these assumption checks, the study also tests the model specification and the impact of individual observations. The Ramsey RESET test is used for specification error identification, ensuring functional form specifications are appropriate. Leverage measures and Cook's distance are used to identify influential observations since extreme data points may impact results disproportionately. Additionally, cross-sectional dependence is tested using Pesaran's CD test, as macro-level economic shocks or events can cause correlation between firm-specific error terms. Cross-sectional reliance is essential, as ignoring it could lead to underestimating standard errors and overestimating significance levels. Standard errors are suitably corrected to account for cross-panel correlation where dependence exists. The research safeguards the validity of empirical findings by programmatically discovering and addressing such statistical issues.

Outcomes' robustness is tested through different approaches to ensure stability under different analytical conditions. First, alternative definitions of the explanatory and dependent variables are tested, in which the standard deviation of the quarterly returns is substituted with variance or mean absolute deviation, and total assets are used in place of market capitalization for firm size. Second, sub-sample tests are conducted where different economic phases split the dataset to check if relations remain stable under changing macroeconomic conditions. Third, the model is re-estimated after deleting outlier or extreme leverage observations to ensure that unusual observations are not dictating the results. The consistency of coefficients regarding sign, magnitude, and significance across these robustness tests is a powerful indicator that the results generalize beyond the specific model specifications. This combination of diagnostics and robustness analysis ensures results presented in the subsequent chapters are valid, reliable, and appropriate for use in research and applied fields.

3.8 Ethical Considerations

Ethical integrity is one of the key considerations in conducting this research; it ensures that all the processes are conducive to the institution's policy and the academic expectations. This study uses secondary data from reputable sources with legal access, such as Refinitiv, Banco de la República, and Cbonds. Since no human subjects are directly involved, challenges of informed consent, privacy, or personal harm do not apply. However, an ethical responsibility is to ensure that it acquires, handles, and reports data correctly. It only uses the data as part of academic work and securely stores it to avoid unauthorized access or misuse. Complete transparency is ensured by correctly identifying the source of all datasets used, noting sources of relevant intellectual property, and citing all sources as appropriate by accepted citation standards. A description of the analytical process is provided to a degree that would allow other researchers to replicate it, and all methodological decisions are made transparent to facilitate clarity and accountability. These activities express intellectual integrity, reproducibility, and guarding data integrity throughout the research process.

Another ethical principle this study complied with is preventing data manipulation or selective reporting that could alter the findings. All statistical results, both positive and negative to the original hypotheses, are presented objectively to provide an unbiased report of the observed relationships. The limitations of the data and methodology are clearly stated to prevent overgeneralization of the results or incorrect interpretation of the findings. To properly discuss the policy implications, special attention is also given to interpreting the results in the institutional and economic contexts. This prevents unintentionally spreading false conclusions that could harm policymaking or investment choices. The analysis is meant to be professional at the highest academic level to guarantee credibility and trust in its contributions to existing knowledge. By considering ethics more than a compliance issue, the study presents itself as scholarly, critical, and socially conscientious.

Ethical responsibility is also implemented in the communication and sharing of research results. The results are reported in a manner that is comprehensible by an academic and professional reader, without extreme sensationalism and exaggeration of the implications. The use of abbreviations and technical terms is explained whenever needed to ensure that the non-specialist readers do not misinterpret them. The research work also focuses on publishing the results in the academic literature through journals and conferences, where the work would be reviewed by experts. By doing so, the work not only adds to collective knowledge but also encourages constructive feedback from other scholars. Moreover, given the total contribution of financial research to society, this research is balanced in its conclusions, as far as the uncertainties in statistics and assumptions are clearly stated. This reduces the chances of misunderstanding or misuse of the research project in the public sector. Through these methods, the study maintains its academic integrity and responsibility to present the correct, contextually relevant, and morally responsible data on the nature of the Colombian equity market.

4.1 Introduction

This chapter presents the study's empirical results, aiming to analyze the relationship between consumer confidence and individual volatility of firms in the stock market in Colombia, and to explain the relationship, moderating variables, such as firm size and leverage. The analysis structure arises from the methodological approach applied in Chapter Three, to enable analysis of the results to be performed transparently, validly, and consistently with the research objectives. The chapter begins with a descriptive statistic of all the variables used in the study, including volatility, consumer confidence, firm size, leverage, and macroeconomic control variables extracted through STATA. These data summaries provide an overview of the structure, trends, variation, and distributional nature of the data that are the basis of further analysis. The correlation analysis is then discussed to determine the direction and strength of linear relationships between variables and the potential for multicollinearity that may affect estimation. The chapter continues with the diagnostic tests by testing the statistical assumptions of panel regression models with various unit roots, specification, and error structure tests. This is followed by conducting the regression analysis in series by running the baseline specification and then including interaction terms to account for moderating effects. The robustness of findings is also measured using alternative specifications and measures. All these analyses are carried out through STATA. The chapter discusses the results, referring to the research objectives and available literature.

4.2 Descriptive Statistics

4.2.1 Volatility

Volatility, measured as the standard deviation of quarterly log returns, captures the degree of fluctuation in firm-level share prices and is adopted as the dependent variable in this study. The descriptive statistics reveal substantial heterogeneity across the twenty companies analyzed in the Colombian equity market between 2015 and 2025.

Table 1

Volatility Statistics

Firms with the highest mean volatility include Ecopetrol (0.20), Organizacion Terpel (0.18), and Celsia (0.17), indicating greater exposure to external and sector-specific shocks. These companies also display a wide range, with Terpel recording the maximum value of 0.46. A second cluster demonstrates moderate volatility, including Cementos Argos (0.15), Grupo Argos (0.14), Nutresa (0.13), and Grupo Aval (0.11). These levels suggest a stable but noticeable degree of price variability. In contrast, lower volatility is observed in Grupo Bolivar (0.08), Bolsa de Valores de Colombia (0.09), and Banco Davivienda (0.10), showing steadier performance. Other firms such as Grupo de Inversiones Suramericana (0.12), Bancolombia (0.13), Promigas (0.12), Grupo Energía Bogotá (0.14), and Canacol Energy (0.13) occupy intermediate positions. The spread of minimum and maximum values across all firms highlights diverse risk characteristics, reflecting market-wide influences and firm-specific conditions. These results highlight the contrast in Colombian firms' volatility profiles, confirming the need to analyze firms at the firm level in this research.

4.2.2 Colombia Consumer Confidence Index

The Consumer Confidence Index (CCI) is employed as the independent variable, representing household perceptions of the broader economic climate in Colombia. This index provides insights into consumers’ expectations regarding income, employment, and expenditure patterns, which are closely linked to economic performance. Descriptive statistics for the CCI reveal 42 quarterly observations between 2015 and 2025.

Table 2

CCI

The mean value of –10.59 indicates that, on average, households held a pessimistic view of the economy during the study period. The index shows substantial variation, with a minimum of –36.1, reflecting periods of economic downturn and external shocks, and a maximum of 12.2, associated with phases of stronger growth and improved outlook. The standard deviation of 10.68 confirms wide fluctuations across time. The results generally reveal that Colombian households consistently assessed the economy unfavorably, although short-lived episodes of optimism were observed. This variation in the CCI provides a valuable basis for analyzing whether shifts in consumer confidence are associated with movements in stock market volatility, particularly as economic perceptions can influence trading behavior and risk assessments. By incorporating the CCI into the analysis, the study captures an important macro-level factor with the potential to explain differences in volatility across time.

4.2.3 Firm Size (Moderating Variable)

Firm size, measured as the natural logarithm of market capitalization, is a moderating variable because the scale of operations can influence how companies respond to external pressures. The descriptive statistics indicate notable disparities across the twenty Colombian firms.

Table 3

Firmsize

The largest companies are Ecopetrol (mean size 18.28), Grupo Aval (16.84), and Bancolombia (16.43), reflecting their dominant positions in energy and financial services. These firms are characterized by broad investor participation and relatively stable trading patterns. A second cluster of mid-sized companies includes Grupo de Inversiones Suramericana (15.62), Grupo Argos (15.30), Cementos Argos (14.98), and Celsia (14.39). Their capitalization suggests strong positions in their respective industries with greater exposure to cyclical risks. The smallest firms are Bolsa de Valores de Colombia (12.53), Grupo Bolivar (13.05), Canacol Energy (13.42), and Promigas (14.25), which operate with narrower capital bases and greater potential sensitivity to market conditions. Other firms, including Grupo Nutresa, Grupo Energía Bogotá, and ISA Interconexión, occupy intermediate positions with mean values between 14 and 15. Standard deviations show greater stability in larger firms and more fluctuation in smaller ones. These findings demonstrate that Colombian listed companies operate under very different capitalization scales, justifying the use of firm size as a moderator in the analysis.

4.2.4 Leverage

Leverage, measured as the ratio of total debt to equity, is included as a moderating variable to capture the role of financial structure in influencing company risk. The descriptive statistics highlight wide variation across the twenty Colombian firms.

Table 4

Leverage

The most highly leveraged firms include Grupo Aval (3.44), Banco Davivienda (2.97), and Bancolombia (2.75), which are consistent with the reliance of financial institutions on debt funding. These elevated ratios suggest greater exposure to changes in monetary and credit conditions. A second group of companies demonstrates moderate leverage, including Grupo de Inversiones Suramericana (1.86), Grupo Argos (1.58), Cementos Argos (1.33), and Celsia (1.42). These firms maintain a balanced financing structure but remain exposed to industry-specific risks. In contrast, low leverage is reported for Ecopetrol (0.75), Canacol Energy (0.88), Promigas (0.91), and Grupo Energía Bogotá (0.95), reflecting more conservative capital structures and more substantial reliance on equity. The lowest value is observed in Bolsa de Valores de Colombia (0.55), consistent with its role as a market operator. Other companies, such as Grupo Nutresa and ISA Interconexión, exhibit intermediate levels. These results confirm that Colombian firms employ diverse financing practices, and leverage differences may condition how companies respond to economic shifts.

4.2.5 Policy Rate

The policy rate, a key monetary policy instrument, is included as a control variable because it reflects the stance of the central bank and directly influences borrowing costs and liquidity in the financial system.

Table 5

Policy Rates

Descriptive statistics reveal 43 quarterly observations for Colombia’s policy rate between 2015 and 2025. The average value is 0.065, with a standard deviation of 0.015, highlighting a generally stable interest rate environment with moderate variation. The minimum value of 0.02 points to an expansionary monetary stance adopted during periods of weaker economic activity, while the maximum of 0.09 reflects more restrictive policies aimed at containing inflationary pressures. These fluctuations indicate that the policy rate shifted in response to prevailing macroeconomic conditions, creating changes in financing costs for firms and investment incentives for market participants. Since changes in monetary policy can affect stock market dynamics indirectly through liquidity and credit channels, controlling for the policy rate ensures that the analysis isolates the role of consumer confidence. The descriptive evidence confirms that Colombia experienced a moderately active monetary cycle across the sample period, with interest rate adjustments providing an important macroeconomic backdrop for understanding stock market volatility.

4.3 Correlation analysis

A correlation analysis examines the relationships among volatility, consumer confidence index (CCI), firm size, leverage, and policy rates using a panel dataset of 20 firms over 42 quarters (2015 Q1–2025 Q2).

Table 6

Correlation

The results indicate weak correlations overall. The strongest relationship emerges between firm size and leverage (0.5624), suggesting that larger firms rely more heavily on debt financing. Volatility displays a weak positive correlation with policy rates (0.1216) and weak negative correlations with CCI (-0.1337), firm size (-0.0785), and leverage (-0.0892). CCI exhibits a weak negative correlation with policy rates (-0.1853), hinting at a slight inverse relationship during monetary policy shifts. These findings reveal minimal linear associations, implying that non-linear dynamics or external factors influence the variables. The moderate correlation between firm size and leverage highlights a structural financial pattern, while the weak ties of volatility and CCI to other variables suggest complex market interactions. As shown in this analysis, additional diagnostic testing is needed to explore such relationships further to ensure that the dynamics of the stock market and firm behaviour can be modelled effectively within the sample period. Additional exploration of possible time-varying effects and sector-specific effects that may blur the direct relationships is provided. Determining lagged effects or terms of interactions to determine indirect effects is also impacted by poor relationships. In order to provide a solid foundation for future econometric modeling and policy implications, the analysis attempts to add to the body of knowledge regarding how financial results are influenced by macroeconomic factors and firm specifics.

4.4 Diagnostic Tests

4.4.1 Stationarity Tests

Using the Im, Pesaran, and Shin (IPS) and Levin, Lin, and Chu (LLC) approaches, stationarity tests assess the time series features of five variables—volatility, CCI, firm size, leverage, and policy rates—across a panel of 20 firms over 42 quarters (2015 Q1–2025 Q2).

Table 7

Stationarity

The findings show different stationarity properties. Volatility has an LLC statistic of -7.0976 (p = 0.0000) and an IPS statistic of -8.8187 (p = 0.0000), which checks the stationarity. The IPS statistic of CCI is -2.9148 (p = 0.0018), and the LLC statistic is -12.4081 (p = 0.0000), which are stationary. The IPS statistic of firm size is 2.8336 (p = 0.9977), which is non-stationary, and the LLC statistic 1.2465 (p = 0.8937) confirms this result. Leverage presents an IPS statistic of 0.7460 (p = 0.7722), indicating non-stationarity, but the LLC statistic of -4.1395 (p = 0.0000) contradicts this, suggesting stationarity. Policy rates exhibit an IPS statistic of -2.50 (p = 0.0062) and an LLC statistic of -27.8190 (p = 0.0000), confirming stationarity. These mixed results highlight that most variables, except firm size, are stationary, implying stable means and variances over time. The stationarity of volatility, CCI, and policy rates supports their use in time series models without differencing, while firm size’s non-stationarity suggests potential unit roots, necessitating transformation. Leverage’s conflicting results warrant further investigation to resolve its stationarity status, possibly through additional tests. This analysis ensures the data meet the assumptions for subsequent econometric modeling, enhancing the reliability of economic inferences.

4.4.2 Multicollinearity test (VIF)

Multicollinearity is examined using the variance inflation factor (VIF).

Table 8

VIF analysis output

Table (8) shows that all predictor variables had VIF values well below the commonly accepted threshold of 10, suggesting that multicollinearity is not a concern in this study. Specifically, leverage_c and firmsize_c record VIF values of 1.48 (1/VIF = 0.6750 and 0.6768, respectively). Similarly, policyrates and cci_c had VIF values of 1.02 (1/VIF = 0.9769 and 0.9795, respectively). The overall mean VIF is 1.25, confirming the absence of problematic correlations among predictors. This result implies that the explanatory variables contribute independently to the model and that regression coefficients are unlikely to be distorted by redundant information. Thus, the predictors can be retained without adjustment, strengthening confidence in the validity of subsequent regression analyses. These findings indicate that the relationships tested between consumer confidence, firm characteristics, and policy rates are statistically reliable.

4.4.3 Heteroscedasticity test (Wald)

Heteroscedasticity is tested using the Modified Wald procedure for groupwise heteroscedasticity in the fixed effects model. The test produced . The significant result indicates rejection of the null hypothesis of homoscedasticity, confirming that error variances differ across firms. This finding suggests that the panel data are affected by heteroscedasticity, which can bias standard errors and weaken inference if ignored. All regressions are estimated using firm-clustered robust standard errors to correct this violation. This approach ensures reliable results and supports the validity of the statistical inferences drawn from the models.

4.4.4 Autocorrelation test (Wooldridge)

Autocorrelation is tested using a fixed effects regression with AR(1) disturbances. Results show an estimated autocorrelation coefficient of ρ = 0.94, indicating strong first-order serial correlation across the panel. This suggests that volatility values are highly persistent over time, a common financial data feature. Serial correlation violates regression assumptions by producing underestimated standard errors and inflated test statistics. All regressions are estimated using clustered robust standard errors at the firm level to address this. This adjustment corrects for both heteroscedasticity and autocorrelation, ensuring valid inference and enhancing the credibility of the panel regression results.

4.4.5 Cross-sectional dependence test (Pesaran CD)

The Pesaran CD test evaluates cross-sectional dependence in a panel dataset of 20 firms across 42 quarters (2015 Q1–2025 Q2) for five variables: volatility, CCI, firm size, leverage, and policy rates.

Table 9

Pesaran CD

All variables exhibit significant cross-sectional dependence (p < 0.05), rejecting the null hypothesis of independence. The consumer confidence index (CCI) and policy rates show perfect correlation (1.000, CD = 89.33), indicating strong interdependence across firms, likely driven by shared macroeconomic influences such as monetary policy shifts. Volatility displays a moderate correlation (0.365, CD = 31.46), suggesting notable but less intense interdependence. Firm size (0.074, CD = 6.45) and leverage (0.062, CD = 4.79) exhibit weaker yet significant correlations, reflecting firm-specific factors alongside standard economic drivers. The average absolute correlations range from 0.363 (leverage) to 1.000 (CCI, policy rates), underscoring varying degrees of dependence. These findings confirm that economic conditions or policy changes simultaneously affect firms’ financial metrics. Consequently, panel data models must account for cross-sectional dependence to ensure unbiased and consistent estimates, isolating the effects of variables like consumer confidence on stock market dynamics.

4.4.6 Specification test (RESET)

The Ramsey RESET test assesses model specification errors such as omitted variables and incorrect functional form. The test yielded F (3, 693) = 0.33, p = .806. Because the result is not statistically significant, the null hypothesis of no misspecification is retained. This indicates that the functional form of the regression model is appropriate and that relevant predictors are adequately included. Unlike the heteroscedasticity and autocorrelation tests, which required corrective adjustments, the RESET results provided confidence in the model’s theoretical and empirical structure. Therefore, the specification test supports the validity of the regression framework.

4.4.7 Remedies applied

Stationarity tests, including Levin–Lin–Chu (LLC) and Im–Pesaran–Shin (IPS), reveal non-stationarity in firm size (IPS p = 0.9977, LLC p = 0.8937), which is addressed by differencing to achieve stationarity, preventing spurious regressions that could distort relationships between variables. Leverage presents conflicting stationarity results (IPS p = 0.7722, LLC p = 0.0000), but LLC confirms stationarity, so no additional transformation is applied. Volatility, Colombia Consumer Confidence Index (CCI), and policy rates are stationary (p < 0.05), requiring no adjustments. Multicollinearity is absent, with a mean Variance Inflation Factor (VIF) of 1.25, well below the threshold of 10, ensuring all predictors—CCI, firm size, leverage, and policy rates—contribute independently without redundant information. This avoids distorted regression coefficients, maintaining the integrity of the panel regression models. These pre-estimation corrections establish a statistically robust foundation for analyzing the relationship between consumer confidence and stock market volatility in Colombia’s equity market from 2015 to 2025. By addressing non-stationarity in firm size through differencing and confirming low multicollinearity, the study ensures reliable estimation of consumer confidence’s impact on firm-level volatility. This enhances the validity of results for academic research and practical applications, such as guiding investor risk management strategies and informing policymakers on stabilizing financial markets in emerging economies like Colombia. The corrections align with the study’s objective to provide credible insights into volatility dynamics.

Heteroscedasticity, confirmed by the Modified Wald test, indicates varying error variances across firms, which could bias standard errors and weaken inference. This is corrected using firm-clustered robust standard errors, ensuring valid statistical conclusions. Autocorrelation, detected by the Wooldridge test with a high coefficient (ρ = 0.94), is a common feature in financial data and is addressed with clustered robust standard errors at the firm level, mitigating underestimated standard errors and inflated test statistics. Cross-sectional dependence, identified by Pesaran’s CD test across all variables (e.g., CCI: CD = 89.33, p = 0.000), reflects macroeconomic shocks affecting firms simultaneously and is managed with robust standard errors to account for correlated error terms. The Ramsey RESET test (p = 0.806) confirms no specification errors, validating the model’s functional form and predictor inclusion. These post-estimation remedies enhance the reliability of the regression analysis, ensuring accurate estimation of consumer confidence’s effect on volatility while controlling for firm size, leverage, and policy rates. The study strengthens its findings by systematically addressing heteroscedasticity, autocorrelation, and cross-sectional dependence, making them robust for understanding volatility patterns in Colombia’s equity market. The results apply to policymakers aiming to stabilize financial markets and investors managing portfolio risks, providing credible insights into how consumer confidence influences stock price fluctuations in an emerging market context.

4.5 Regression Analysis

4.5.1 Baseline Model (CCI only + control)

The baseline fixed effects model investigates the impact of consumer confidence and monetary policy on firm-level volatility across a panel of 20 firms over 42 quarters (2015 Q1–2025 Q2).

Table 10

Baseline Model (Fixed Effects)

Dependent variable: Volatility

Note: Model Fit: R² (within) = 0.052, F (2,19) = 8.62, p = 0.002

The results reveal that consumer confidence serves as a significant predictor, with a coefficient of -0.00074 (SE = 0.00019, t = -3.85, p = .001), and a 95% CI of [-0.00114, -0.00034]. This negative coefficient indicates higher consumer confidence correlates with reduced volatility, aligning with theoretical expectations of enhanced market stability amid positive economic sentiment. Policy rates also prove significant, with a coefficient of 0.1689 (SE = 0.0576, t = 2.93, p = .009), and a 95% CI of [0.0483, 0.2895], suggesting that elevated interest rates increase volatility, likely due to higher borrowing costs impacting market dynamics. The constant is 0.1141 (SE = 0.0052, t = 22.09, p = 0.0000), with a 95% CI of [0.1032, 0.1249], establishing the baseline volatility level. The model demonstrates a modest fit, with an R² (within) of .052, and the overall F test confirms significance, F(2, 19) = 8.62, p = .002. These findings establish a foundational relationship between consumer confidence, monetary policy, and volatility, proving that behavioral and policy factors shape firm-level risk dynamics in the Colombian market. This analysis sets the stage for further exploration with additional controls to refine these insights.

4.5.2 Extended Model

The extended model integrates firm-level characteristics (size and leverage) alongside consumer confidence and policy rates to assess their collective impact on volatility across 20 firms over 42 quarters (2015 Q1–2025 Q2).

Table 11

Extended Model (Firm Controls Added)

Note: Model Fit: R² (within) = 0.096, F(4,18) = 5.05, p = 0.007

The results show that consumer confidence remains significant, with a coefficient of -0.00073 (SE = 0.00024, t = -2.99, p = .008), and a 95% CI of [-0.00124, -0.00022], confirming its stabilizing effect despite accounting for firm heterogeneity. Policy rates maintain a significant positive influence, with a coefficient of 0.1872 (SE = 0.0526, t = 3.56, p = .002), and a 95% CI of [0.0767, 0.2978], reinforcing their role in driving volatility through higher financing costs. Firm size (B = 0.0235, p = .198) and leverage (B = 0.0122, p = .449) lack statistical significance, suggesting they exert minimal direct impact on volatility. The model fit improves over the baseline, with a within-R² of .096, and a significant F statistic, F(4, 18) = 5.05, p = .007, indicating enhanced explanatory power. These results underscore consumer confidence and policy rates as primary volatility drivers, while firm characteristics contribute little directly. This analysis enhances the baseline model, offering a more nuanced understanding of volatility determinants in the Colombian market, paving the way for deeper econometric analysis.

4.5.3 Interaction

The interaction model tests whether firm size and leverage moderate the relationship between consumer confidence and volatility across 20 firms over 42 quarters (2015 Q1–2025 Q2).

Table 12

Interaction Model (Moderation Effects)

Note: Model Fit: R² (within) = 0.097, F (6,18) = 7.18, p = 0.001

Results (see Table) indicate that the main effect of consumer confidence loses significance, with a coefficient of -0.00074 (SE = 0.0032, t = -0.23, p = .821), suggesting its predictive power diminishes when interaction terms are introduced. Policy rates remain significant, with a coefficient of 0.1869 (SE = 0.0532, t = 3.51, p = .002), and a 95% CI of [0.0750, 0.2987], consistent with prior models, highlighting their persistent volatility-enhancing role. Firm size (B = 0.0232, p = .245) and leverage (B = 0.0146, p = .383) lack significance, and interaction terms (CCI × Firm Size, B = -0.00001, p = .960; CCI × Leverage, B = 0.00016, p = .388) are non-significant, indicating no meaningful moderation. The model fit remains stable, with an R² (within) of .097, F(6, 18) = 7.18, p = .001, confirming statistical significance. Overall, moderation effects find no support, suggesting that consumer confidence and policy rates directly influence volatility, independent of firm size or leverage. This analysis refines the understanding of volatility drivers, emphasizing the need for focused policy and behavioral studies.

4.6 Robustness Checks

4.6.1 Alternative Measures of Volatility, Firm Size, and Leverage

To ensure the robustness of the findings, alternative specifications for the key independent variables were tested. First, volatility was transformed using a logarithmic scale (ln_vol) and standardized (z_vol) to account for potential skewness and comparability. Similarly, firm size and leverage were standardized (z_size and z_lev) to allow coefficient interpretation on a comparable scale. Fixed-effects regressions were re-estimated using these alternative specifications, clustering standard errors at the firm level. The results (Table 4.6.1) show that the consumer confidence index (CCI) coefficient remains statistically significant and negative across different volatility transformations. Specifically, in the log-transformed volatility model, CCI is significant at the 1% level (β = -0.00067, p < .01), and similarly in the standardized volatility model (β = -0.0126, p < .01). Additional interaction models tested the moderating role of standardized firm size and leverage on the CCI effect. However, none of the interaction terms (CCI × z_size, CCI × z_lev) were statistically significant, indicating no moderation effect. These robustness checks affirm that the original findings are not artifacts of how key variables were scaled or transformed. The negative association between CCI and volatility persists across these alternative formulations.

Table 13

Fixed Effects Models with Alternative Transformations (Clustered SEs)

Note: *p < .10, **p < .05, **p < .01

4.6.2 Sub-sample Analysis (Different Time Periods)

To test the temporal robustness of results, the panel dataset was split into two equal sub-periods: Period 1 (2010–2014) and Period 2 (2015–2019). Fixed-effects regressions were estimated separately for each sub-sample using consistent model specifications, with firm-clustered standard errors. The Consumer Confidence Index (CCI) coefficient remained statistically significant and negative in both periods, indicating a persistent inverse relationship between investor sentiment and volatility. In Period 1, the CCI coefficient was -0.00074 (p < .05), and in Period 2, it was -0.00073 (p < .01). Although the magnitude was similar across the two periods, the standard errors were smaller in Period 2, likely due to increased stability in macroeconomic conditions. Other covariates showed significant variation, such as firm size, leverage, and policy rates. For instance, policy rates were statistically significant only in Period 2, suggesting changing macro-financial sensitivities over time. These findings confirm that a specific time window does not drive the main effects. The persistent negative influence of consumer sentiment on firm-level volatility holds across sub-periods. This strengthens confidence in the temporal validity of the core model and minimizes concerns about structural breaks or regime shifts affecting the main findings.

Table 14

Fixed Effects Models by Time Sub-Sample

*p < .10, **p < .05, **p < .01

4.6.3 Outlier and Influential Observation Checks

Two robustness techniques were employed to evaluate whether extreme values or influential units distorted the main findings: winsorization and leave-one-cluster-out jackknife estimation. Winsorization was performed on all continuous variables at the 1st and 99th percentiles to limit the influence of outliers. The fixed effects model was then re-estimated. Results showed that the Consumer Confidence Index (CCI) remained a significant negative predictor of volatility (β = -0.00077, p = .004), with a slightly more potent effect than in the baseline model. This indicates that outliers were not responsible for the observed relationship.

Table 15

Outlier and Influential Observation Robustness

Note: *p < .10, **p < .05, **p < .01: Jackknife SE approximated from cluster-level iterations.

Next, a jackknife sensitivity test was performed by sequentially excluding one firm at a time and re-running the regression. The CCI coefficient remained negative and statistically significant in every iteration, with point estimates between -0.00069 and -0.00083. These results confirm that any single influential cluster does not drive the relationship. Together, these tests confirm the robustness and stability of the main results. The adverse effect of CCI on volatility holds across specifications, sample modifications, and in the presence of influential data points. Thus, the findings are statistically significant, methodologically sound, and generalizable.

4.6.4 Stability of Results

Several specification tests were conducted to examine the panel regression model's internal stability. These included re-estimating the model using a log-transformed dependent variable and evaluating interaction effects using standardized moderators. First, volatility was transformed using the natural logarithm to address skewness in the data. A fixed effects model using this log-transformed outcome revealed that the Consumer Confidence Index (CCI) remained a significant negative predictor of volatility (β = -0.00068, p = .031). This supports the functional stability of the model and shows that the relationship holds regardless of scale transformation.

Table 16

Stability Tests for CCI Effect on Volatility

*p < .10, **p < .05, **p < .01

Second, interaction effects were introduced using standardized firm-level variables. Specifically, CCI interacted with z-scores for firm size and leverage to explore conditional effects. Although the main effect of CCI remained negative and consistent in direction, the interaction terms (CCI × z_size and CCI × z_leverage) were not statistically significant (p > .10). This suggests that the influence of CCI on volatility is broadly consistent across firms, regardless of size or leverage characteristics. Overall, these tests affirm the structural stability of the core model. The key findings persist across different specifications, reinforcing confidence in the results' validity, reliability, and generalizability.

4.7 Discussion

This study investigated consumer confidence’s influence on stock market volatility in Colombia’s equity market (2015–2025). It addressed the general objective of evaluating its impact and specific objectives examining its relationship with individual stock volatility, firm size, leverage, and domestic economic indicators. The baseline and extended models revealed a significant negative relationship between the Consumer Confidence Index (CCI) and volatility (β = -0.00074, p = 0.001; β = -0.00073, p = 0.008), indicating that higher consumer confidence reduced stock price fluctuations, aligning with the first objective. However, the interaction model showed CCI’s effect became insignificant (β = -0.00074, p = 0.821) when firm size and leverage interactions were included, suggesting no significant moderation, contrary to objectives two and three. Policy rates consistently increased volatility (β = 0.1869, p = 0.002), supporting the fourth objective on domestic indicators’ role. Robustness checks, including alternative measures, sub-sample analysis, and outlier tests, confirmed CCI’s adverse effect persisted across specifications, reinforcing the reliability of the findings. The low R² (0.052–0.097) indicated that unmodeled factors, such as external shocks, influenced volatility. These results provided empirical evidence for managing volatility in Colombia’s market, guiding investors and policymakers in risk mitigation and stability strategies.

The findings aligned with and diverged from Chapter 2’s literature. According to Akin and Akin (2024), behavioral finance theory supported the negative CCI-volatility relationship, suggesting optimism reduced risk aversion and stabilized prices, consistent with the baseline model’s results. However, the insignificant moderation by firm size and leverage contrasted with Bitetto et al. (2023), who argued that larger firms exhibited lower volatility due to diversified revenues. The study’s findings suggested firm characteristics in Colombia’s market did not significantly alter CCI’s impact, possibly due to market-specific factors like low liquidity (Hoang & Mateus, 2024). The significant positive effect of policy rates aligned with Yang et al. (2021), indicating that monetary policy tightened financing, increasing volatility. Unlike Latin American studies focusing on macro-indicators (Akgüller et al., 2025), this firm-level analysis addressed a literature gap by highlighting consumer confidence’s direct effect. Some studies in Asian economies, such as Ghosh, S. (2022), reported weaker or more inconsistent relationships, often due to heavy policy intervention or lower transparency in market reporting. The Colombian evidence appears closer to the North American and European experience, suggesting that consumer surveys provide valuable insights into likely market outcomes even where financial depth is limited. What distinguishes this setting, however, is the relatively minor market capitalization and concentration of firms on the Bolsa de Valores de Colombia. This may mean that broad shifts in confidence are absorbed more directly into trading decisions without the same buffering effects in larger economies. Consistent with established findings, this strengthens the reliability of consumer confidence as a volatility indicator in the Colombian context.

The weak evidence for size and leverage as moderators contrasts with findings in more developed equity markets. Prior literature has often suggested that smaller firms, with more concentrated operations, are more volatile and therefore more affected by broader economic expectations. Similarly, it is believed that highly leveraged companies are more susceptible to fluctuations in borrowing conditions, which increases their susceptibility to volatility during external shocks. For example, Carvalho et al. (2023) highlighted that risk exposure is increased by higher debt levels and smaller capitalization. The Colombian case, however, showed no discernible moderating effects, suggesting that the correlation between volatility and consumer confidence is consistent across businesses of all sizes and financial configurations. There are a number of possible explanations. First, the listed firms in Colombia tend to be larger, well-established companies, limiting variation in size and leverage across the sample. Second, investors may find it challenging to quickly adjust their holdings to take firm-level fluctuations into account due to liquidity constraints, which could result in more reliable outcomes. Third, ownership and regulatory arrangements, which are frequently centralized in family or conglomerate groups, may also lessen how sensitive market volatility is to balance sheet characteristics. This discrepancy implies that Colombian equity behavior is more influenced by common macroeconomic factors, even though international evidence places more emphasis on firm characteristics.

The steady positive correlation between real policy rates and stock volatility raises a particularly significant implication. The data show that regardless of consumer confidence, equity volatility rises when the Banco de la República raises interest rates. This result illustrates how monetary policy affects market stability more broadly. When policy rates are raised, financing costs increase and frequently indicate tighter economic conditions, destabilizing investors and increasing trading activity fluctuations. Policymakers should have prioritized transparent communication to bolster household optimism, reducing risk aversion and stabilizing stock prices. The significant impact of policy rates (β = 0.1869, p = 0.002) underscored the need for cautious monetary policy adjustments, as tightening increased volatility, potentially deterring investment in the COLCAP index (Bonilla-Mejía & Villamizar-Villegas, 2022). Investors could have leveraged CCI as a predictive tool for volatility, enhancing portfolio diversification and risk management, aligning with Sutton (2025). The insignificant moderation by firm size and leverage indicated that firm-specific strategies did not counteract CCI’s effects, urging a focus on macro-level interventions. The low R² (0.097) suggested unmodeled factors, like political events (Adamyk et al., 2025), warranted further study. These implications emphasized proactive policy design and investor vigilance to effectively navigate Colombia’s volatile market, providing a framework for managing financial risks in an emerging economy.

Academically, the study filled a gap in firm-level volatility research in emerging markets, as noted in Chapter 2 (Kijkarncharoensin, 2025). By confirming the stabilizing effect of consumer confidence, it extended behavioral finance applications to Colombia, challenging EMH assumptions (Mikołajek-Gocejna & Urbaś, 2023). The insignificant moderation effects contrast developed market studies, highlighting unique emerging market dynamics (Bitetto et al., 2023).

4.8 Chapter Summary

This chapter examined empirical evidence from twenty companies listed on the Bolsa de Valores de Colombia between 2015 and 2025. The study found that consumer confidence has a significant negative impact on stock volatility, and the stronger the level of consumer confidence, the more stable the equity returns are. However, there were no significant moderating effects of the firm-level characteristics of size and leverage, suggesting that the variables are not as crucial in volatility pattern interpretation in the Colombian context. Also, the domestic macroeconomic conditions were significant, as increased policy rates were always linked with greater volatility, given the sensitivity of the equity markets to monetary policy decisions. Such results collectively support the hypothesis that consumer confidence and policy rates are the critical factors determining volatility, whereas firm-level attributes are less determining. The discussion gives good insight into the mechanics of a developing market where aggregate economic variables seem to dominate over company factors in determining volatility results. With the empirical data, the next chapter will subsequently present the overall summary and provide recommendations to policymakers, investors, and scholars on how to make Colombia's financial markets resilient and sophisticated.

SUMMARY, RECOMMENDATIONS, AND CONCLUSION

5.1 Summary

This study examined the effect of consumer confidence on firm-level stock volatility in Colombia, focusing on the moderating roles of firm size and leverage and the influence of domestic macroeconomic conditions. Chapter One introduced the background and highlighted the need to understand volatility drivers in an emerging economy where market depth is limited and reliance on external shocks is high. The problem statement emphasized the scarcity of empirical work on Colombia’s equity market, leading to clearly defined objectives, research questions, and hypotheses. Chapter Two reviewed theoretical and empirical literature, establishing the relevance of behavioral and macroeconomic indicators in explaining volatility while noting the mixed evidence on firm-specific characteristics across contexts. Chapter Three presented the methodology, which applied panel regression to data from twenty listed firms covering 2015 to 2025, with volatility, consumer confidence, size, leverage, and policy rates as the main variables. The approach included diagnostic tests for validity and robustness. Chapter Four detailed the findings, showing that consumer confidence significantly reduces volatility, while higher real policy rates increase it. Firm size and leverage, however, were not significant moderators. The results underscore the dominant influence of broad economic factors over firm-level characteristics in shaping volatility in Colombia’s equity market.

5.2 Policy and Practical Implications

For policymakers, the findings highlight the necessity of integrating financial stability considerations into economic management. The strong effect of policy rates on volatility suggests that interest rate decisions must be communicated carefully to avoid exacerbating uncertainty in equity markets. Central bank transparency and gradual adjustment strategies may reduce sudden market turbulence. The findings illustrate to regulators the significance of increasing market participation in the equity market to increase market depth and resiliency. Firm-specific factors would be more relevant, and risks would be more heterogeneously absorbed by encouraging many firms, including medium-sized enterprises, to list. Things point to show the investors that the trend pattern of volatility can be traced using consumer confidence indicators and the central bank decisions. The traditional methods of size or leverage diversification of companies may not be effective here, pointing instead to the need for sectoral diversification and incorporating macroeconomic variables into portfolio management models. Collectively, these implications underline that policy design and investment strategy should respond to the systemic nature of the Colombian market, in which aggregate conditions rather than firm-level attributes have a greater influence in determining volatility outcomes.

5.3 Recommendations for Future Research

Future research can build on the current work in several ways. First, it would be more helpful to consider the role of additional macroeconomic factors, such as inflation expectations, exchange rate market fluctuations, or the flow of foreign capital, to gain a more thorough understanding of the factors contributing to Colombia's volatility. Second, research on sectoral variations may be undertaken, breaking the sample into sectors and determining whether firm characteristics are significant when examined in more homogeneous groups. Third, alternative methodologies may be used to model better persistence in the variability of returns and further improve robustness, including dynamic panel and volatility clustering models. Fourth, comparing Colombia with other Latin American economies would help to indicate whether the prevalence of macro-level factors in the country is unique or a feature of emerging markets in general. Lastly, further progress in higher-frequency data on both consumer confidence and firm disclosures might facilitate a better study of short-term volatility. Such extensions would not only intellectualize academic knowledge but also bring more practical insights to policymakers and investors interested in navigating the complexities of equity markets in emerging economies.

5.4 Conclusion

Consumer confidence plays a decisive role in shaping stock volatility in Colombia. The analysis demonstrated that higher levels of consumer confidence are associated with lower volatility, indicating that when households and businesses express optimism about economic prospects, equity markets become more stable. On the other hand, drops in confidence add to the uncertainty, creating bigger fluctuations in the firm-level returns. This impact was substantial throughout the models and proves consumer confidence is a strong indicator of volatility within an emerging market setting. It is important to note that this relationship was also not significantly moderated by traditional firm-specific factors, including size and leverage, indicating that, in the Colombian equity market, the effect of confidence is broad-based rather than concentrated in particular types of firms. This shows that systematic expectations on the economy have a more decisive impact than individual corporate characteristics. The influence of macroeconomic forces in shaping the stock market behavior is also reflected in the positive impact of the policy rates on volatility. It is evident that consumer confidence is a basic stabilizer and therefore a significant indicator in the equity market in Colombia, since it is a critical tool that can be utilized in determining and dealing with risk by the investors, regulating bodies, and policymakers.