Asymmetric Impact of Active Management on the Performance of ESG Funds

Abstract

This paper investigates the asymmetric impact of fund active management style on the performance of ESG funds. Unlike conventional measures of synchronicity, we propose new measures that capture the asymmetric patterns in a fund’s management style in upside and downside market conditions. Our data includes 170 equity funds that are identified as socially responsible, with a period spanning from 2010 to 2022. Our proposed methodology allows us to capture the asymmetric patterns in the fund management styles under different market conditions while mitigating the challenge of outliers, which is crucial when assessing funds’ active management activities. We find that while ESG funds promote sustainability, their active management is only beneficial during periods of market downturns. Our results are robust after controlling for different funds characteristics, for several active management proxies, and across various model specifications. This paper thus provides crucial guidelines for fund managers since it shows that their success is greatly influenced by their time-varying skills and management style in changing market conditions. Our findings incentivize ESG fund managers to pursue information acquisition activities during market downturns, as these activities improve market informational efficiency while aligning with their sustainability goals.

1. Introduction

Environmental, Social, and Governance (ESG) investments have generated significant interest within the past decade. As of 2023, the total global assets under management in sustainable investments has reached $30.3 trillion (Global Sustainable Investment Alliance 2023). This is also complemented by a substantial increase in the number of mutual funds dedicated to ESG investing. MSCI also reported that sustainable ETFs have seen record inflows of $75 billion in 2020, tripling those seen in 2019 (MSCI 2020). The ESG funds have also seen significant inflows as investors seek investment options which are aligned with their values. For instance, sustainable funds attracted net flows of $326.7 billion globally in 2020 (Morningstar 2023). Likewise, managers of traditional mutual funds have also increasingly integrated ESG factors within their investment analyses, decision-making, and portfolio construction.

The increased interest of mutual funds in ESG investing has attracted significant scrutiny in academic literature (Bauer et al. 2007; Mollet and Ziegler 2014; Lins et al. 2017; Bofinger et al. 2022). Most of this literature aims to assess the impact and the value of ESG investments. The findings, however, are far from conclusive. Some papers find an insignificant relationship between ESG investing and fund performance (Renneboog et al. 2008; Shanthirathna et al. 2023). Other studies suggest that investing in socially responsible instruments can contribute to value destruction (Bauer et al. 2007; Borgers et al. 2015; El Ghoul and Karoui 2017), partially due to the low levels of diversification and partially due to the limited investment opportunities in the portfolio allocation process (Renneboog et al. 2008). Socially responsible investments have also been shown to incur costs, mainly monitoring costs that investors pay for obtaining non-financial gains (Benson and Humphrey 2008), which can contribute towards their underperformance. In contrast, several studies demonstrate the superior performance of ESG funds compared to their non-sustainable counterparts, thereby challenging the myth that ESG investing sacrifices returns. For instance, Gil-Bazo et al. (2010) along with Nofsinger and Varma (2014) find that the benefits of ESG investing surpass its costs. Cortez et al. (2009) complement these findings by reporting that the screening process used by ESG funds—a central aspect of socially responsible investing—can produce higher returns due to the presence of better information quality.

The decisions, strategies, and approaches taken by fund managers play a crucial role in determining their performance compared to their peers. For instance, the management style of fund managers can significantly influence their performance. Given the contradictory evidence on the value of different management styles (Berk and Van Binsbergen 2015; Chen and Scholtens 2018; El Ghoul and Karoui 2022; Klinkowska and Zhao 2023), it is worth revisiting the impact of active management on fund performance. This paper aims to explore this relationship within the context of ESG funds. However, several challenges are associated with the current active management proxies employed in the literature. For example, proxies for active management such as the degree of industrial concentration of Kacperczyk et al. (2005), the R-squared metrics utilized by Amihud and Goyenko (2013) and those employed by Huij and Derwall (2011) do not adequately capture the sensitivity to dataset outliers and to market condition changes. We argue that these measurements of active management may be distorted by outliers, producing inaccurate assessments of a fund manager’s skill. The performance of fund managers under varying market conditions is also not sufficiently reflected by the current active management proxies used in the literature. Managers’ skills and capacity to outperform can be affected and vary by whether the market is bullish, bearish, or undergoing other unique conditions. Current active management proxies used in the literature often fail to account for the asymmetric nature of market movements, leading to incomplete assessments of fund performance. We argue that a more nuanced methodology is needed to evaluate the effectiveness of active fund management. By proposing new measures of active management, our study aims to provide a clearer understanding of the conditions under which a fund’s activeness can enhance its performance. In this context, our study makes significant contributions to the literature by addressing the limitations of existing active management measures and highlighting the importance of market conditions in evaluating fund performance.

The novelty of our paper thus consists of proposing new active management proxies that capture the asymmetric patterns in funds’ management style in upside and downside market conditions while addressing the challenges associated with the influence of outliers. We then assess the impact of our proposed active management proxies on the ESG funds’ performance. The decision to actively manage socially responsible investment portfolios presents several benefits. However, we believe that its effectiveness during upside market conditions remains dubious. Several papers argue that the relative performance of active fund management is not persistent over time, and this especially holds when considering performance net of fees (Fama and French 2010; Glode 2011; Nofsinger and Varma 2014). The costs associated with active funds can be significant. Most of these costs are due to screening mechanisms adopted by fund managers. Screening costs are expected to reduce the asset universe and expose investors to underdiversification. This generates risky investment portfolios that may underperform their respective benchmarks in normal period conditions (Renneboog et al. 2008; Gangi and Varrone 2018). We believe that this can influence the underperformance of active management in upside market conditions.

Conversely, in downside market conditions, we argue that investors often pay heightened attention to the performance of their investments (Cao et al. 2011; Guiso et al. 2018). This is especially relevant and crucial to the ESG funds that are perceived as safe and resilient investment vehicles. We thus believe that ESG fund managers are faced with the responsibility of demonstrating their skills and expertise during crisis periods by actively managing their portfolios and proactively responding to market challenges, which can positively influence their performance. We also argue that active fund management may perform better during turbulent periods for several additional reasons. First, corporate managers tend to withhold bad information more than good news during downturns, which may lead to less information being breached to the market. The relatively high information opaqueness during these times thereby provides a better opportunity for fund managers to display their skills and expertise by actively managing their portfolios (Lantushenko and Nelling 2021). Second, active funds have the flexibility to adjust their portfolios in response to market conditions, which can help them mitigate risk during downturns. Third, market downturns can create pricing inefficiencies, such as panic selling or overreaction to negative news, which can lead to mispriced securities. Therefore, there may be valuable investment opportunities in undervalued assets. Active funds are in a better position to identify mispriced securities that can offer opportunities for outperformance.

Consistently with the above arguments, our findings suggest that managers who adopt active management strategies in downside market conditions create positive and significant performance, while in upside market conditions, active management is shown to be detrimental to funds’ performance. In other words, adopting passive management strategies in good states of the economy enhances ESG funds’ performance. Interestingly, our results also show that the impact of active management on the performance of funds can be misleading when employing conventional active management measures that do not account for market conditions states. Our findings are robust after controlling for different funds characteristics, for several active management proxies and across various model specifications. Furthermore, we also show that several fund-specific characteristics moderate the relationship between management style and fund performance.

This paper presents several contributions. First, the methodology employed allows us to define the management styles adopted by funds as a function of upside/downside market conditions. Our methodology allows us to capture the asymmetric patterns in the fund management styles in different market conditions while mitigating the challenge of outliers that is crucial when assessing funds’ active management. Our findings provide a compelling reason for fund managers to increase their efforts in acquiring information during market downturns. By focusing on acquiring valuable insights within ESG funds, managers can improve market efficiency and reduce mispricing during turbulent periods, aligning with their sustainability objectives. We structure the remainder of this paper in the following way: Section 2 presents the literature and the theoretical framework. Section 3 describes the data and presents the methodology. Section 4 and Section 5 document the results. The paper concludes with Section 6.

2. Literature Review and Theoretical Framework

2.1. Literature Review

The aim of this paper is to document the impact of fund management style on the performance of ESG funds. The management style of fund managers can be broadly divided into two categories—active management and passive management. Passive fund management operates under the premise that markets are efficient in the sense that they incorporate all available information. As a result, passive funds do not seek to outperform the market. Passive management seeks to match the performance of a specific market index, rather than striving to beat it. To achieve this, passive managers create a portfolio that closely replicates the composition of the target index. Instead of frequently trading, they focus on maintaining a portfolio that aligns with the index’s performance. This approach typically involves lower fees since it demands less research and trading activity, resulting in reduced transaction costs and lower management fees. Conversely, active fund management operates under the assumption that markets may take time to incorporate information. Therefore, it is possible to outperform the market by developing expertise in identifying mispriced securities and by using the market-timing strategies. The latter strategies involve the ability of fund managers to implement investment decisions based on predictions of future market movements, effectively timing the buying and selling of assets to maximize returns. This approach comes with higher costs due to increased fees. Active management involves greater risk, as it depends on the success of the investment choices made. In contrast, passive management’s risk is tied to the potential underperformance of the market index itself. This fundamental difference in management style has led many researchers to investigate the relative value of active management in comparison to passive fund management (Muñoz et al. 2015; Chen and Scholtens 2018; El Ghoul and Karoui 2022; Klinkowska and Zhao 2023). Some studies find that active managers fail to generate excess returns after accounting for funds’ expenses, thereby exposing investors to losses (Fama and French 2010; Chen and Scholtens 2018). In contrast, other papers suggest that the strategies and skills of active fund managers, such as stock picking, sector rotation, and market timing, generate positive and significant gains for investors (Cremers and Petajisto 2009; Berk and Van Binsbergen 2015). These funds tend to achieve significantly high persistent performance before and after expenses.

We argue that the value of fund management style is time-varying and depends on the state of the market. As highlighted above, the decision to actively manage a socially responsible investment portfolio presents several benefits. However, we believe that its effectiveness during upside market conditions is dubious. Several studies argue that the value of active fund management is not persistent over time, particularly when considering the fund performance net of fees (Fama and French 2010; Glode 2011; Nofsinger and Varma 2014). The expenses associated with active ESG funds can be substantial, mainly stemming from screening mechanisms implemented by fund managers. These screening expenses reduce the asset universe, potentially exposing funds to inadequate diversification. In the same line, Bofinger et al. (2022) find that a “sustainability trap” is present for active fund managers. The authors provide evidence of a positive correlation between ESG rating of funds and overpricing. This indicates that active managers who increase the sustainability levels of their portfolios face the risk of converting their portfolios into a mediocre-skill fund. We believe that this can influence the underperformance of active management in upside market conditions.

We propose that there are higher incentives for active management when markets are going down. Our arguments are consistent with Moskowitz (2000), who provides evidence that while actively managed mutual funds tend to perform poorly, they become more valuable to investors during turbulent times as they deliver superior performance. Glode (2011) also supports these arguments by developing a model in which a fund manager chooses to focus their efforts during periods where investors’ marginal utility of consumption is higher. The author argues that active funds’ high returns during more difficult periods may not be properly acknowledged by the strict assessment of the fund’s overall performance.

In addition to this, the high information opaqueness and the increased information asymmetry during downturn periods result in reduced information disclosure to the market. This provides better opportunities for fund managers to display their skills and expertise through active portfolio management (Lantushenko and Nelling 2021). Hence, we argue in this paper that ESG active funds are in a better position to identify mispriced securities that can offer opportunities for outperformance during turbulent periods.

Another crucial issue we address in this paper is the shortcomings of existing active management proxies in the literature. Primarily, these measures are often subject to estimation errors and are not adequately adjusted to capture the sensitivity to market condition changes. Amihud and Goyenko (2013), as well as Huij and Derwall (2011), are well known for using the R-squared predictor, which is widely recognized as an active management measure of fund performance. This measure, by its nature, is sensitive to deviations in data, leading to problems where outliers can disproportionately influence regression results. To address this, Amihud and Goyenko (2013) censor the extreme 0.5% tails of the data distribution of the estimated R-squared to eliminate outliers and closet indexers (where managers claim an active management style but invest passively). Additionally, the R-squared measure focuses on the extent to which fund returns deviate from the benchmark, which presents some drawbacks. Mainly, the choice of an appropriate benchmark for the fund’s activeness can be biased, leading to potential misestimation of active management measures.

The choice of an appropriate benchmark and the restricted accessibility to fund holdings data remain as primary challenges of different active management proxies employed in the literature. For instance, Active Share is another well-known measure of active management introduced by Cremers and Petajisto (2009). The authors use weight-based portfolio composition to measure divergence from the benchmark and conclude that an increased active management style enhances fund performance. A portfolio’s divergence is measured by quantifying the extent to which a portfolio’s returns deviates from its benchmark returns, better known as tracking error. In a similar vein, Kacperczyk et al. (2005) assert that managers who restrict their holdings to specific industries compared to their benchmarks tend to perform better for the overall sample period.

Conversely, Doshi et al. (2015) presented a novel measure of active management that captures the absolute difference between the value weights and the actual weights held by a fund, averaged across its holdings. High-active funds are shown to outperform low-active funds by 2.63% per year. While this measure is empirically efficient, it considers volatility timing with the Generalized Method of Moments methodology employed by Ferson and Mo (2016). Therefore, this measure becomes highly context-dependent. In another study, Shankar (2007) affirms the outperformance of active management over passive management through the proxy of portfolio turnover. However, the use of turnover as a measure of active management presents some drawbacks, since higher turnover implies higher transaction costs and tax liabilities, which can offset the benefits of active management. Therefore, in this paper, we aim to propose new active management proxies that capture the asymmetric patterns in funds’ management styles under both upside and downside market conditions while addressing the challenges associated with the influence of outliers. Our approach allows for a more precise assessment of fund managers’ abilities to navigate different market conditions, offering valuable insights that can enhance investment strategies. Furthermore, we apply these proxies to assess the performance of ESG funds, contributing to the growing body of research on the effectiveness of sustainable investing.

2.2. Theoretical Framework

The current literature highlights the fact that active mutual fund managers engage in a complex process of acquiring and analyzing private information to identify undervalued securities with the aim of outperforming the market. The challenge lies in earning sufficient abnormal returns to compensate for transaction costs and information-acquisition activities fees. This aligns with the asset pricing rational expectations equilibrium models that consider asymmetrically informed investors (Grossman and Stiglitz 1980; Easley and O’hara 2004). The pursuit of active management for sophisticated investors such as ESG funds—which are considered informed participants within these theoretical frameworks—remains a challenging balancing act between generating alpha and managing expenses. The theoretical basis of our paper is grounded in the Grossman and Stiglitz (1980) paradox, which questions the Efficient Market Hypothesis (EMH). Grossman and Stiglitz argue that if markets were efficient, investors would be less willing to seek additional information, as it would not offer any excess returns. Nevertheless, the process of collecting information and identifying mispriced securities suggests that markets present different levels of inefficiency.

To address this issue, we explore the theoretical framework underlying the models of rational expectation equilibrium, with a particular focus on the model proposed by Grossman and Stiglitz (1980). The authors proved that there is a positive relationship between the amount of information acquired by uninformed investors and the share of informed investors in the market. This is also true for the level of accuracy of the information they have acquired. It is worth noting, however, that the amount of information acquired by these uninformed investors decreases as investors become more risk-averse and with the increase in market noise.

We highlight the importance of this model when it comes to evaluating the level of information efficiency in the stock market in periods of market crisis, as there is a marked increase in risk aversion predominantly due to fear and uncertainty (Cao et al. 2011). Many studies show that in the aftermath of a destabilizing event, traders tend to be more risk averse. This was highlighted after the 2008 Global Financial Crisis (Guiso et al. 2018) and following the events of the COVID-19 pandemic (Fassas 2020). In such situations, obtaining and analyzing information becomes increasingly difficult (Giannetti and Laeven 2012), which typically lowers the accuracy of the information available to investors, assuming other factors remain constant. According to the model developed by Grossman and Stiglitz, if the proportion of informed investors remains unchanged, equilibrium prices will demonstrate reduced informational efficiency. Active ESG funds, who are considered informed investors within the framework of Grossman and Stiglitz (1980), can thus have a substantial influence over both their performance and their financial markets during periods of financial turmoil. During such periods characterized by heightened market noise and diminished informational efficiency in financial markets (Anagnostidis et al. 2016), it becomes crucial for fund managers to trade actively based on their informational advantage to efficiently navigate financial crises.

One might question why ESG fund managers would invest in information acquisition during financial crises. Fund managers could be worried about whether they can achieve enough profits to justify the increased costs of gathering and analyzing information. ESG funds in particular often have higher fees due to additional expenses related to ESG factors, such as screening and compliance costs. Despite these concerns, the heightened noise trading during crises means that prices do not effectively reflect the insights of sophisticated investors to less-informed ones. This dynamic maintains an information advantage for well-informed fund managers. Consequently, these managers are motivated to devote time and resources to acquire information, as they believe it will help them identify valuable investment opportunities and generate returns sufficient to cover their costs. Adding to that, pursuing information-based active management allows ESG fund managers to align their portfolios with the investor’s expectations and will enhance long-term performance. Investors may shift their interest to ESG funds during financial crises, seeking stability and resilience (Cheng et al. 2014; Lins et al. 2017), as these funds are perceived to be better equipped to navigate economic downturns. ESG funds can also be valuable for investors since they emphasize transparency and accountability. ESG fund managers may thus have incentives to actively engage in information activities to enhance and preserve investor’s confidence, especially in challenging market conditions.

Our paper aims to assess the impact of ESG funds’ active management activities on their performance in different market conditions. We believe that ESG funds are often perceived as stable, resilient, and trustworthy investment vehicles, offering strong incentives for fund managers to actively acquire information. Following the Grossman and Stiglitz framework, we argue that active information acquisition enables ESG funds to seize unique market opportunities and achieve superior performance that can offset the higher costs associated with the funds’ information acquisition activities. Based on the Grossman and Stiglitz model, we also hypothesize that the increased market noise and diminished informational efficiency during financial crises can enhance the value of private information held by ESG fund managers, leading to better performance relative to uninformed investors. In addition, we posit that the incentives for ESG fund managers to acquire and utilize private information during crises can be driven by the need to meet investor’s expectations and sustain investor confidence. Therefore, we propose the following hypotheses:

H1. 

Active management of ESG funds is beneficial during periods of market growth.

H2. 

Active management of ESG funds is beneficial during periods of market turmoil.

3. Data and Methodology

3.1. Sample

We chose a sample that encompasses all mutual funds from the United States that are identified as socially responsible by the Forum for Sustainable and Responsible Investment (USSIF). We retrieved the data from the CRSP Survivor-Bias-Free Mutual Fund Database. We cover around 170 equity funds for the period between the third quarter of 2010 and the third quarter of 2022.

3.2. Methodology and Definition of Variables

This paper proposes new measures of active fund management that are functions of market conditions. We aim to assess the synchronicity measures that capture the degree to which fund returns move in alignment with market returns. Our measures allow us to capture the asymmetric patterns in the management style of funds as follows:

$$\mathrm{UPASYNCH}=-\left({\displaystyle \frac{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{s} {\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}}_{\mathrm{j},\mathrm{t}}>0|{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}}_{\mathrm{m},\mathrm{t}}>0}{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{s} {\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}}_{\mathrm{m},\mathrm{t}}>0}}\right)$$

(1)

$$\mathrm{DOWNASYNCH}=-\left({\displaystyle \frac{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{s} {\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}}_{\mathrm{j},\mathrm{t}}<0|{\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}}_{\mathrm{m},\mathrm{t}}<0}{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{s} {\mathrm{R}\mathrm{E}\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}}_{\mathrm{m},\mathrm{t}}<0}}\right)$$

(2)

To obtain the upside asynchronicity (UPASYNCH) variable, we perform the following steps: First, we measure the proportion of times the fund and the market returns were jointly positive during the last 12 months. We then multiply the result by −1 to obtain the upside asynchronicity variable for each fund. In the same line, to obtain the downside asynchronicity (DOWNASYNCH) variable, we start by measuring the proportion of times the fund and the market returns are jointly negative during the last 12 months. We then multiply the result by −1 to obtain the downside asynchronicity variable for each fund. Higher values of UPASYNCH (DOWNASYNCH) indicate more active management during periods of market growth (decline). We use the previous 250 days (the approximate number of trading days within a year) to estimate the proxies of active management. In other words, we are calculating these measures based on the data from that year. The same approach applies to the computation of alpha. It is worth noting that our variables are not affected by the presence of outliers. This allows us to mitigate the impact of extreme values as well as to obtain robust and reliable proxies of active management.

These variables are used in the following regression to test the impact of active management on the performance of ESG funds:

$$\begin{gathered} {\mathrm{A}\mathrm{L}\mathrm{P}\mathrm{H}\mathrm{A}}_{\mathrm{j},\mathrm{t}}=\mathsf{\alpha }+{\mathsf{\beta }}_1\left({\mathrm{U}\mathrm{P}\mathrm{A}\mathrm{S}\mathrm{Y}\mathrm{N}\mathrm{C}\mathrm{H}}_{\mathrm{j},\mathrm{t}}\right)+{\mathsf{\beta }}_2\left({\mathrm{D}\mathrm{O}\mathrm{W}\mathrm{N}\mathrm{S}\mathrm{A}\mathrm{Y}\mathrm{N}\mathrm{C}\mathrm{H}}_{\mathrm{j},\mathrm{t}}\right)+{\mathsf{\beta }}_3\left({\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{\mathrm{j},\mathrm{t}}\right)+{\mathsf{\beta }}_4\left({\mathrm{F}\mathrm{L}\mathrm{O}\mathrm{W}}_{\mathrm{j},\mathrm{t}}\right)+ \\ {\mathsf{\beta }}_5\left({\mathrm{E}\mathrm{X}\mathrm{P}\mathrm{E}\mathrm{N}\mathrm{S}\mathrm{E}}_{\mathrm{j},\mathrm{t}}\right)+{\mathsf{\beta }}_6\left({\mathrm{A}\mathrm{G}\mathrm{E}}_{\mathrm{j},\mathrm{t}}\right)+{\mathsf{\beta }}_7\left({\mathrm{T}\mathrm{U}\mathrm{R}\mathrm{N}\mathrm{O}\mathrm{V}\mathrm{E}\mathrm{R}}_{\mathrm{j},\mathrm{t}}\right)+{\displaystyle {\sum }_{\mathrm{Y}=1}^{\mathrm{N}-1}}{\mathsf{\gamma }}_{\mathrm{Y}}\left(\mathrm{Y}\mathrm{D}\mathrm{U}\mathrm{M}\right)+{\mathsf{\epsilon }}_{\mathrm{t}} \end{gathered}$$

(3)

In the above regression model, the performance of each fund j is measured by the abnormal return that is captured by the alpha (ALPHA) of the Fama–French three factor model. The Fama–French three factor model is estimated by using the previous 36-day window interval. Our main independent variables of interest—UPASYNCH and DOWNASYNCH—measure the level of active management. Furthermore, we also include the funds characteristics, such as the logarithm of total assets (SIZE), the percentage change in total assets (FLOW), the expense ratio (EXPENSE), the logarithm of fund’s age (AGE), and the turnover rate (TURNOVER), as the control variables. The choice of control variables is based on prior literature (Amihud and Goyenko 2013; El Ghoul and Karoui 2022). For the purpose of completeness, we also include year dummies (YDUM) to control for the year-specific affects. The t-values from this regression are computed using robust standard errors. These robust standard errors address concerns related to both linearity and heteroskedasticity, ensuring that our statistical inferences remain valid. We used yearly data for all our variables. This alignment ensures consistency in the timing of our measurements across all variables, facilitating a coherent analysis. See Appendix A for the definition of variables.

3.3. Summary Statistics

Table 1. Descriptive statistics.

Variables	25th Percentile	Mean	Median	75th Percentile	Standard Deviation	Observations.
DOWNASYNCH	−0.2602	−0.2095	−0.2383	−0.1726	0.0736	6252
UPASYNCH	−0.3268	−0.2700	−0.3041	−0.2434	0.0834	6252
EXPENSE	0.0065	0.0102	0.0095	0.0125	0.0050	6252
FLOW	−0.0217	0.0544	0.0332	0.0964	0.2142	6252
SIZE	3.1441	4.3601	4.5239	5.5789	1.8652	6252
AGE	1.7351	2.2363	2.3969	2.9085	0.8949	6252
TURNOVER	0.2100	0.6955	0.3900	0.8500	0.8485	6252

Note: see Appendix A for definition of variables.

documents the summary statistics of the variables used in this paper. The average value of DOWNSYNCH is −0.2095, which indicates that a little less than 21 percent of the time, the fund and the market returns move together in the negative direction. The average value of UPSYNCH is −0.2700, which suggests that 27 percent of the time, fund and market returns move together in the positive direction.

Table 2. Correlation matrix.

No.	Variables	1	2	3	4	5	6	7	VIF
1	DOWNASYNCH	1.0000							4.56
2	UPASYNCH	0.8771 ***	1.0000						4.52
3	EXPENSE	−0.1078 ***	−0.0853 ***	1.0000					1.32
4	FLOW	0.0382 ***	−0.0287 **	−0.0607 ***	1.0000				1.16
5	SIZE	−0.0093	−0.0201	−0.3666 ***	−0.1408 ***	1.0000			1.85
6	AGE	−0.0885 ***	−0.0935 ***	0.0688 ***	−0.3388 ***	0.5430 ***	1.0000		1.77
7	TURNOVER	0.5047 ***	0.4965 ***	−0.0113	0.0219 *	−0.0064	−0.0425 ***	1.0000	1.37

Note: The symbols *, **, *** correspond to p < 0.1, p < 0.05, p < 0.01, respectively. See Appendix A for definition of variables.

documents the VIF values along with the correlation between variables used in this paper. The table shows that multicollinearity between control variables is low, thereby allowing us to include all of them together in a regression.

4. Results

Table 3. Relationship between active management and fund performance.

Variables	Model (1)	Model (2)	Model (3)	Model (4)
UPASYNCH	−0.0003		−0.0123 ***
	(−0.8666)		(−18.2741)
DOWNASYNCH		0.0027 ***	0.0155 ***
		(7.9062)	(19.7471)
TOTALSYNCH				0.0005 ***
				(3.3036)
AGE	0.0001 *	0.0001 **	0.0001 **	0.0001 **
	(1.8070)	(2.5450)	(2.0431)	(2.1695)
EXPENSE	−0.0014	0.0013	0.0012	−0.0001
	(−0.3387)	(0.3259)	(0.2911)	(−0.0338)
TURNOVER	0.0001 ***	−0.0001 **	−0.0001	0.0001
	(3.7338)	(−2.2764)	(−0.7178)	(0.8624)
FLOW	0.0012 ***	0.0012 ***	0.0010 ***	0.0012 ***
	(8.2044)	(8.3078)	(7.4667)	(8.2915)
SIZE	0.0001	0.0001	0.0001	0.0001
	(1.4196)	(1.4289)	(1.5478)	(1.4176)
Year Dummies	Yes	Yes	Yes	Yes
Observations	6252	6252	6252	6252
R-Squared	0.7842	0.7867	0.8001	0.7846
F-Values	2273.60 ***	2408.78 ***	3666.45 ***	2287.44 ***

Note: This table displays the relationship between active management and fund performance, whereas the dependent variable is the mutual funds’ net alpha estimated from Fama–French three factor model. Our main independent variables of interest measure the extent of active management: (UPASYNCH), (DOWNASYNCH), and total synchronicity (TOTALSYNCH). Furthermore, we also include the funds characteristics such as the logarithm of total assets (SIZE), the percentage change in total assets (FLOW), the expense ratio (EXPENSE), the logarithm of fund’s age (AGE), and the turnover rate (TURNOVER) as control variables. The t-values based on the pooled OLS regression with robust standard errors are shown in parenthesis. The symbols *, **, *** correspond to p < 0.1, p < 0.05, p < 0.01, respectively. See Appendix A for definition of variables.

documents the coefficient estimates of Equation (3). The findings show that the coefficient estimate of DOWNASYNCH is significantly positive. It is an indication that funds which shift to a more active management in response to market declines perform better than similar funds that do not shift to a more active management. Also, we find a significantly negative coefficient estimate of UPASYNCH. This indicates that funds which shift to a more active management in response to upside market conditions underperform other funds. These results support the arguments presented in this paper by highlighting that active fund management is only valuable when markets are declining. Our findings thus reject our hypothesis H1 and confirm our hypothesis H2. Furthermore, in order to elaborate that the findings based on conventional proxies of active management, such as the measure of Amihud and Goyenko (2013), do not show the true value of active management, we modify Equation (3) by replacing DOWNASYNCH and UPASYNCH with a variable (TOTALSYNCH) that does not define active fund management with respect to the market conditions. This variable represents the proportion of times when the fund returns and the market returns are jointly either positive or negative. The findings suggest that TOTALSYNCH has a positive impact on fund performance. This finding is misleading because as we have already shown, active ESG fund management is only valuable when markets are declining and less beneficial during periods of market growth. Our results thus highlight the asymmetric impact of fund management based on the market status.

5. Additional Tests

For robustness checks, we conduct additional tests to verify the consistency of our findings across various estimation strategies and to examine how different managerial characteristics of funds might influence our results. These tests ensure that our results are not sensitive to the specific technique used and enhance the credibility and comprehensiveness of our conclusions.

5.1. Relationship between Active Management and Fund Performance: Alternate Estimation Strategies

In order to test the reliability of our findings, we re-estimate Equation (3) by using alternate estimation strategies. For example, the findings in Model (1) are based on the pooled OLS regression with clustered standard errors at the fund-level. This strategy allows for correlated residuals at the fund-level. With this strategy, it is possible to produce unbiased standard errors, if the residuals are uncorrelated across funds. Model (2) adjusts for cross-sectional correlation and heteroskedasticity by using the Fama–MacBeth regression. Model (3) makes use of the fact that the data is organized as a panel. The findings of this model are based on panel regression with fixed effects. The last two models are based on the pooled OLS regression with robust standard errors. In Model (4), we use one-period ahead value of ALPHA to control for reverse causality. In Model (5), we aim to address some of the concerns regarding omitted variable bias by introducing lagged value of ALPHA as an independent variable. The findings in

Table 4. Relationship between active management and fund performance: alternate estimation strategies.

Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
UPASYNCH	−0.0123 ***	−0.0024 **	−0.0138 ***	−0.0048 ***	−0.0089 ***
	(−15.7187)	(−2.6555)	(−15.3459)	(−10.4256)	(−16.6087)
DOWNASYNCH	0.0155 ***	0.0039 ***	0.0172 ***	0.0065 ***	0.0111 ***
	(17.6616)	(3.6708)	(19.8129)	(12.9747)	(17.103)
ALPHA (Lagged)					0.5274 ***
					(25.0537)
AGE	0.0001 **	0.0001	0.0001	0.0001	0.0001 **
	(2.4881)	(1.2988)	(1.2163)	(0.3886)	(2.4205)
EXPENSE	0.0012	−0.0047	0.0092	−0.0059 *	0.0017
	(0.4114)	(−1.1174)	(0.8658)	(−1.8884)	(0.4473)
TURNOVER	−0.0001	−0.0001	−0.0001	0.0001	−0.0001
	(−0.6946)	(−0.7050)	(−0.3721)	(0.0682)	(−1.0813)
FLOW	0.0010 ***	0.0002 **	0.0010 ***	−0.0001	0.0010 ***
	(6.4671)	(2.1131)	(6.3358)	(−1.2846)	(7.5861)
SIZE	0.0001 **	0.0001	0.0001 *	−0.0001	0.0001
	(2.4634)	(−1.0118)	(1.9388)	(−0.4656)	(0.8974)
Year Dummies	Yes	No	Yes	Yes	Yes
Observations	6252	6252	6252	6098	6098
R-Squared	0.8001	0.2191	0.7973	0.9107	0.8575

Note: This table displays the relationship between active management and fund performance, whereas the dependent variable is the mutual funds’ net alpha estimated from the Fama–French three factors model. Our main independent variables of interest measure the extent of active management in upside and downside market conditions: (UPASYNCH) and (DOWNASYNCH). We also include the funds characteristics such as the logarithm of total assets (SIZE), the percentage change in total assets (FLOW), the expense ratio (EXPENSE), the logarithm of fund’s age (AGE), and the turnover rate (TURNOVER) as control variables. The t-values are shown in parenthesis. Model (1) is based on the pooled OLS regression with clustered standard errors at the fund-level, Model (2) is based on the Fama–MacBeth regression, Model (3) is based on the panel regression with fixed effects, and Model (4) and Model (5) are based on pooled OLS regression with robust standard errors. The symbols *, **, *** correspond to p < 0.1, p < 0.05, p < 0.01, respectively. See Appendix A for definition of variables.

confirm the arguments presented in this paper by showing that active fund management is only valuable when markets are not doing well. In periods of market advances, active fund management does not add value.

5.2. The Link between Active Management and Fund Performance: The Impact of Fund Characteristics

In order to test the impact of fund characteristics on our findings, we modify Equation (3) by introducing interaction of UPSAYNCH and DOWNASYNCH with fund characteristics as independent variables. If these interaction variables are significant, it is an indication that the relationship between active management and fund performance is sensitive to fund characteristics. The results of this analysis are presented in

Table 5. Effect of fund characteristics on the relationship between active management and fund performance.

Variables	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)
UPASYNCH	−0.0055 ***	−0.0132 ***	−0.0150 ***	−0.0123 ***	−0.0096 ***
	(−3.9717)	(−9.6961)	(−17.8764)	(−17.1522)	(−6.1184)
DOWNASYNCH	0.0072 ***	0.0171 ***	0.0182 ***	0.0158 ***	0.0119 ***
	(4.3598)	(10.493)	(18.2534)	(19.4492)	(6.6752)
AGE*UPASYNCH	−0.0031 ***
	(−5.2540)
AGE*DOWNASYNCH	0.0038 ***
	(5.4473)
EXPENSE*UPASYNCH		0.0759
		(0.6377)
EXPENSE*DOWNASYNCH		−0.1530
		(−1.1284)
TURNOVER*UPASYNCH			0.0035 ***
			(5.9397)
TURNOVER*DOWNASYNCH			−0.0034 ***
			(−4.5647)
FLOW*UPASYNCH				−0.0013
				(−0.3372)
FLOW*DOWNASYNCH				−0.0061
				(−1.2322)
SIZE*UPASYNCH					−0.0007 **
					(−2.0098)
SIZE*DOWNASYNCH					0.0009 **
					(2.2567)
AGE	0.0001	0.0001 **	0.0001	0.0001 **	0.0001 **
	(0.1277)	(2.0596)	(1.5427)	(2.1250)	(2.0549)
EXPENSE	0.0025	−0.0112	0.0029	0.0012	0.0013
	(0.6349)	(−0.7411)	(0.6964)	(0.2949)	(0.3287)
TURNOVER	0.0001	0.0001	0.0002 ***	0.0001	0.0001
	(0.0174)	(−0.8087)	(3.3648)	(−0.4968)	(−0.6685)
FLOW	0.0010 ***	0.0010 ***	0.0010 ***	−0.0005 **	0.0010 ***
	(7.4092)	(7.4218)	(7.4536)	(−2.4539)	(7.5286)
SIZE	0.0001 *	0.0001	0.0001	0.0001	0.0001
	(1.7870)	(1.4877)	(1.4943)	(1.5204)	(0.6520)
Year Dummies	Yes	Yes	Yes	Yes	Yes
Observations	6252	6252	6252	6252	6252
R-Squared	0.8016	0.8002	0.8012	0.802	0.8004

Note: This table displays the effect of fund characteristics on the relationship between active management and fund performance, where the dependent variable is the mutual funds’ net alpha estimated from Fama–French three factors model. Our main independent variables of interest measure the extent of active management in upside and downside market conditions: (UPASYNCH) and (DOWNASYNCH). We include interactions of UPASYNCH and DOWNASYNCH with funds characteristics such as the logarithm of total assets (SIZE), the percentage change in total assets (FLOW), the expense ratio (EXPENSE), the logarithm of fund’s age (AGE), and the turnover rate (TURNOVER) as independent variables. The t-values based on the pooled OLS regression with robust standard errors are shown in parenthesis. The symbols *, **, *** correspond to p < 0.1, p < 0.05, p < 0.01, respectively. See Appendix A for definition of variables.

. The findings suggest that the relationship between UPASYNCH and fund performance is more pronounced (more negative) for older funds and larger funds and less pronounced (less negative) for funds with higher turnover. We also show that older funds and larger funds exhibit a stronger relationship between DOWNASYNCH and fund performance and funds with higher turnover exhibit a weaker relationship between these two variables.

Our results in Table 4 and Table 5 provide interesting findings related to the overall impact of managerial characteristics on fund performance. These tables indicate a positive and significant impact of age on fund performance in two models and no relation in three models in Table 4 while showing significance in three models in Table 5. These findings are consistent with the mixed results in the literature. For example, Ferreira et al. (2013) found no relationship between the age of U.S. mutual funds and their performance but noted that newer non-U.S. funds tend to outperform older ones. Similarly, Farid and Wahba (2022) discovered that older funds underperform compared to younger ones. Conversely, Alvi and Rehan (2020) argue that older funds, with more experience, can generate favorable returns. In addition, our results indicate a positive significant impact of size on fund performance in two of the five models in Table 4 and one model in Table 5. Research on this topic has shown varied outcomes. Rehman and Baloch (2016) and Ferreira et al. (2013) found that larger non-U.S. funds achieve better risk-adjusted performance, whereas smaller U.S. funds perform better. Conversely, Farid and Wahba (2022) found a significant negative relationship, suggesting that larger funds underperform due to diseconomies of scale. Junaeni (2022) found no significant effect, emphasizing the importance of the investment manager’s strategy over fund size itself.

Regarding fund expense ratio, our results show a negative effect on fund performance in one model and insignificant results in the others. Studies on this relationship also reveal varied conclusions. Ferreira et al. (2013) found a negative but statistically insignificant correlation for U.S. funds. Rehman and Baloch (2016) suggested that a higher expense ratio might positively impact performance, while Junaeni (2022) found no significant effect, attributing differences to the manager’s strategy. Alvi and Rehan (2020) suggested that higher expenditures could enhance performance due to better resources.

Furthermore, our results show no significant effect of turnover on fund performance. Research on asset turnover and fund performance shows mixed results. Rehman and Baloch (2016) found a positive effect, while Junaeni (2022) reported no significant impact, emphasizing market timing over turnover ratio. Ramesh and Dhume (2014) noted that higher turnover increases costs, affecting performance. Finally, our results indicate a positive significant impact of flows on fund performance in most models. Studies have also shown mixed findings. Ferreira et al. (2013) found no significant correlation for U.S. funds but observed a better performance with higher inflows in non-U.S. funds. Junaeni (2022) reported that increased fund cash flow leads to better performance. Conversely, Ramesh and Dhume (2014) found that higher inflows can diminish performance, particularly for smaller funds. Overall, while some studies suggest higher fund flows can improve performance, others indicate potential negative impacts, especially for smaller funds.

6. Conclusions

In this paper, we argue that there are asymmetric incentives for ESG fund managers to adopt active or passive management strategies depending on the status of the market, which can significantly impact their performance. The objective of this paper is to address several challenges associated with the current active management proxies employed in the literature. For instance, active management measures such as the degree of industrial concentration used by Kacperczyk et al. (2005), the R-squared metrics utilized by Amihud and Goyenko (2013), and those employed by Huij and Derwall (2011) fail to adequately capture sensitivity to dataset outliers and shifts in market conditions. We contend that these active management metrics can lead to inaccurate evaluations of a fund manager’s skill. Moreover, the current proxies do not sufficiently account for the performance of fund managers under different market conditions. We argue that a manager’s skills and ability to outperform can vary significantly depending on whether the market is bullish, bearish, or experiencing other unique conditions. Therefore, we propose measures of funds’ active management that capture the asynchronicity between the fund and market returns in different market conditions. Our proposed measures are not affected by the presence of outliers. This allows us to mitigate the impact of extreme values and obtain robust proxies of active management. Our results show that ESG funds have greater incentives to adopt active management during the periods of market decline. These incentives are relatively low when markets are advancing. Consistent with these arguments, this paper demonstrates that funds shifting to more active management in response to market declines perform better than otherwise similar funds that do not shift to more active management. The paper also shows that funds shifting to more passive management in response to market advances perform better than otherwise similar funds. Interestingly, our findings also demonstrate that conventional proxy of active management generates misleading conclusions regarding the effect of management strategies on fund performance.

Our study makes significant contributions to the literature by addressing the limitations of existing active management proxies and highlighting the importance of market conditions in evaluating fund performance. This paper not only advances theoretical understanding but also provides practical guidelines for ESG fund managers seeking to optimize their management strategies in alignment with sustainability goals. It also provides crucial guidelines for fund managers, as it shows that their success is greatly influenced by their time-varying skills and management style in changing market conditions. Our results can thus incentivize fund managers to increase their efforts in acquiring information during times of market turmoil, as active management can enhance fund performance and improve the informational efficiency of financial markets. Furthermore, this research has practical implications for investors’ asset allocation and risk management since it is essential for investors to assess whether the potential benefits of active management, such as alpha generation (generation of excess returns relative to a fund’s benchmark index) and risk reduction, surpass the higher implied costs in different market conditions.