Assessing Mutual Fund Performance in China: A Sector Weight-Based Approach

Abstract

In many financial markets across the globe, full historical position disclosure is not required of mutual funds, or it is subject to prolonged delays, often due to regulatory restrictions. This makes measuring fund manager performance based upon the stock-picking and market-timing skills from past literatures impossible. This study introduces a new methodology utilizing sector weight analysis to estimate the stock-picking and market timing skills of 198 Chinese equity mutual fund managers. Within-sample predictions confirm that the new measures are robust and reliably identify fund managers who outperform their peers, suggesting that this method may be useful in other institutional settings where the full historical position of funds is unavailable. Fund managers with lower stock picking or market timing skills are more likely to improve their skills in the following period, which suggests that manager skills develop and change over time. Finally, our analysis reveals that fund managers with higher stock picking skills are significantly less likely to be replaced, thereby enjoying greater job security.

1. Introduction

The role of investment in the economy is to reallocate capital from low- to high-efficiency businesses [1]. However, due to the specialized knowledge required for investment decision making, many investors delegate their investment decisions to mutual fund managers [2]. The ability of these managers to make informed investment selections directly impacts economic outcomes, influencing investors’ willingness to invest [3]. Current mutual fund-related research predominantly measures managers’ abilities in terms of stock picking [4] and market timing [5,6]. Additionally, fund-level factors and personal characteristics of managers also contribute to their overall performance [7].

When investors choose managers to delegate their investment behaviors, they care about the current performance, the persistence of the manager’s past performance, and the performance volatilities [8]. Unfortunately, the persistence evidence is closer to mid-term than long-term [9]. The persistence is explained by management fees and transaction costs rather than performance. Some research points out that only top-performing funds show some persistence [10]. When the focus is on emerging markets, the story may differ since the market volatility in emerging markets is large, and the information is often not transparent. Suppose that passive investment may also experience large volatilities and the investors believe that fund managers may have an information advantage. In that case, it makes it more likely for the investors to seek professional delegation [11]. In another aspect, managers quickly react to strategies and style changes when they realize they are underperforming other fund managers. They could adopt new ideas by mimicking or learning from other fund managers. Such behavior could make persistence difficult for the outperformed funds [12].

The Chinese financial market has experienced rapid growth in recent years and has emerged as a distinctive representative among emerging markets. In particular, the number of mutual funds in China has seen a significant surge since 2016. As in other emerging markets’ financial systems, information asymmetry remains prevalent, resulting in greater market volatility compared to developed markets [13]. Given the ongoing economic development and rapid industrial updates, investment, capital reallocation, and efficiency improvements are especially crucial in the Chinese financial market, making it an ideal candidate for study.

The Chinese investors do not show smart fund selection skills [14]. This blind selection could be attributed to low-performance persistence, as performance volatility makes selection difficult [15]. Evidence suggests that the smaller fund outperforms competitors by generating a more significant positive future alpha [16]. The fund flow–performance analysis shows that outperforming funds do not significantly attract later inflows [17]. When focusing only on the bull market, the fund flow performance becomes significant and positive [18].

This paper contributes to the current academic research in the following ways. Firstly, we propose a new approach to assess mutual fund managers’ stock picking and market timing abilities using fund market sector allocation. This method can be applied even in cases where the full historical stock holding the position information of the funds is unavailable, as in China and many other financial markets around the world. We validate the robustness of our approach through in-sample tests of whether higher scores correctly identify fund managers who outperform their peers. Secondly, while the existing body of research includes methods for exploring the stock-picking and market-timing abilities of fund managers, there a lack of research comparing how those skills evolve over time. We delve into this aspect and further connect managers’ past performance and ability with their current ability and the likelihood of improving this ability. Finally, we investigate how fund managers’ stock-picking and market-timing abilities affect job security and the related question of how replacement of a fund manager affects fund performance in the Chinese market.

The remainder of this paper is organized as follows. Section 1 provides the background for this study, while Section 2 reviews the existing literature and forms several hypotheses based on established theory. Section 3 explains the data and empirical methodology employed to analyze these hypotheses, and Section 4 discusses the empirical results. Finally, Section 5 concludes the paper.

2. Literature Review and Hypotheses

The body of literature concerning the performance of mutual funds has seen extensive development, yielding a range of research findings that present a mixed picture. Early studies failed to provide substantial support for the notion that mutual fund managers possess distinctive skills, emphasizing instead that only underperforming managers exhibit any semblance of performance persistence [19]. However, as we delve into shorter return sample intervals, more recent evidence has surfaced indicating that fund managers may indeed possess timing abilities [20].

In scenarios where managers enjoy an information advantage, they tend to assign greater weight and allocate more funds to equities in which they have heightened confidence. This information advantage has the potential to significantly enhance fund performance [21,22]. It is noteworthy that fund managers often find it more manageable to oversee funds with smaller asset sizes. Recent findings suggest that short-term performance persistence is evident in smaller funds over a six-month period [23]. This observation aligns with the concept of an information advantage and the idea that smaller funds are more amenable to diversification, consequently leading to improved performance [24].

Previous research has generally pointed to a negative relationship between a fund’s portfolio turnover rate and its performance [25]. This phenomenon may be attributed to the increased trading costs incurred with excessive trading activities [26].

Considering the redemption and past performance perspective, empirical evidence underscores the notion that redemptions and fund outflows following periods of poor performance are more pronounced than inflows following strong performance periods [27].

It is important to note that fund performance exhibits a positive correlation with the fees assessed [28]. Funds that incorporate front and back loads tend to have lower expenses and yield higher risk-adjusted returns [29]. Inflows are typically driven by superior relative performance, whereas outflows are predominantly tied to absolute poor performance [30].

2.1. New Measures of Stock-Picking and Market-Timing Skill

Market-timing and stock-picking abilities are crucial for mutual fund managers to outperform passive investment strategies. Managers who possess strong timing skills may choose stocks with better liquidity, especially when they prioritize timing decisions [31]. Self-confidence can drive highly skilled managers to be more active in their investment timing approaches [32]. The impact of timing ability on returns is particularly significant for both top- and bottom-performing managers [33]. In addition to manager attributes, the attributes of the fund, including its clientele, may matter. Larger fund families with lower portfolio turnover tend to exhibit better timing ability [34], and managers of funds with more profit-seeking investors as clients demonstrate better timing abilities [35].

Market timing decisions can be influenced by local market characteristics as well, making individualized and context-specific assessments necessary [36]. More experienced fund managers tend to accumulate greater timing experience over time [37], but the volatile nature of emerging markets presents challenges for mutual fund managers when it comes to making effective timing decisions [12]. Banking-related mutual funds may possess an information advantage, which contributes to their superior timing ability [38]. Moreover, both mutual fund managers and private fund managers engage in market timing to secure deals with high value [39]. Fund managers with prior foreign investment experience or a track record of successful management are more likely to attract fund subscriptions [40].

Another critical ability required for mutual fund managers to outperform passive investment benchmarks is stock picking [41]. Stocks selected by high-performing mutual fund managers can serve as valuable investment signals for other investors [42]. Corporate governance practices of target firms may play a significant role in managers’ ability to pick the right stocks. Moreover, firms that are frequently selected by funds tend to develop into acquirer firms [43]. The level of outperformance exhibited by a fund can indicate the persistence of its superior stock picking ability [44]. In the Chinese IPO market, the unobservable variable of mutual fund managers’ expectations shows a correlation with significant returns. The picking ability of managers can be influenced by fund and market characteristics [45], with mutual funds being more inclined to allocate investments to pure-play firms rather than conglomerates [46]. Interestingly, fund managers who start their careers during a recession may exhibit more conservative behavior by picking defensive stocks over cyclical stocks [47].

In a seminal study on measuring the stock picking and timing abilities of mutual fund managers, Kacperczyk et al. (2014) introduced a measure of the ability of fund managers to pick stocks well, particularly important in booms, and to time the market, which is crucial during recessions [48]. Kacperczyk et al.’s (2014) stock-picking ability measure measures how a fund manager’s stock selection for the fund relative to the market comoves with the systematic component of the stock’s return. The timing ability measure measures how a fund manager’s holding of each stock relative to the market comoves with the idiosyncratic component of the stock return. To utilize Kacperczyk et al.’s (2014) stock-picking and timing ability measures, however, complete historical data on the full holdings of the mutual funds in question are required. In the Chinese market, funds are required to report their top 10 holdings, but historical data on the full fund position is not disclosed. Despite the more limited disclosure in the Chinese market, funds in China do report their historical sector asset allocations. The regulatory authority classifies the market into 19 sectors, and each fund reports its allocation among these sectors on a quarterly basis.

In our proposed measure, we first calculate a sector beta by performing regression analysis to determine the quarterly beta of all stocks within each sector. Then, using the stock capitalization on the final day of the quarter as the weight, the stock-weighted sector beta is calculated. The sum of all weighted betas for the stocks within each sector provides 19 unique sector betas for each quarter. This sector beta represents the standard market correlation of stocks in any of the 19 sectors i in fund j, based on the Capital Asset Pricing Model (CAPM).

Using these calculated sector betas, Equation (1) is used to measure the stock-picking ability of fund managers and Equation (2) to assess fund managers’ market-timing abilities. These equations are nearly identical to Kacperczyk et al.’s (2014) [48] measures, but our modified measure calculates sector stock picking and sector market timing rather than individual stock picking and individual stock market timing.

$$Pickin{g}_t^j={{\displaystyle \sum }}_{i=1}^{N^j}(w_{i,t}^j-w_{i,t}^m)\left(R_{t+1}^i-{\beta }_{i,j}{R}_{t+1}^m\right)$$

(1)

$$Timin{g}_t^j={{\displaystyle \sum }}_{i=1}^{N^j}(w_{i,t}^j-w_{i,t}^m)\left({\beta }_{i,j}{R}_{t+1}^m\right)$$

(2)

where:

$w_{i,t}^j$ represents the portfolio weight of sector i in fund j at time t, calculated as the fraction of the fund j’s total assets held in sector i at the start of time t;

$w_{i,t}^m$ represents the portfolio weight of sector i in the market benchmark m, calculated as the fraction of the selected market benchmark m’s, the China Security Index 300 (CSI300), and the total assets held in sector i at the start of time t;

$R_{t+1}^i$ represents the average return of the stocks in sector i at time t + 1;

$R_{t+1}^m$ represents the market benchmark m’s return at time t + 1;

${\beta }_{i,j}$ measures the covariance of the returns of stocks in sector i, $R_{t+1}^i$, with the market benchmark return, $R_{t+1}^m$, divided by the variance of the market benchmark return.

Note that Beta ${\beta }_{i,j}$ in Equations (1) and (2) above is the standard market correlation based on the Capital Asset Pricing Model (CAPM), but here, it is the correlation of sector i in fund j rather than the correlation of an individual stock. The product of ${\beta }_{i,j}{R}_{t+1}^m$ measures the systematic component of sector i’s returns.

The measures of Picking and Timing calculated in Equations (1) and (2) are expressed in quarterly returns. Picking measures how a fund’s holdings of each sector, relative to the market, comove with the idiosyncratic component of the stock return. Timing measures how a fund’s holdings of each sector, relative to the manager’s belief about future market performance, comove with the systematic component of the sector stock return.

To test the robustness of our new measure, we conduct an in-sample test of the ability of these new measures of stock-picking and market-timing skills to identify fund managers that outperform their peers in absolute terms. Considering information asymmetries that have been documented to exist between firms in the Chinese financial market [49,50] and evidence showing that Chinese mutual funds hold insider information [51], we also hypothesize that higher stock-picking and market-timing skills raise risk-adjusted relative fund performance.

Hypothesis 1 (H1).

Fund managers with higher stock-picking or market-timing skill ability outperform their peers in both absolute and risk-adjusted relative returns.

2.2. Learning Stock Picking and Market Timing Skills

Are stock-picking and market-timing skills innate talents that one is born with? Or are they skills that change and develop with experience?

Kacperczyk et al. (2014) [48] demonstrated that the stock-picking and market-timing skills exercised by fund managers vary with the market environment. In particular, Kacperczyk et al. (2014) [48] provided convincing empirical evidence that fund managers tend to focus more on stock picking during booms and to apply more market-timing skill during recessions. Skilled managers alternate between timing and picking to consistently outperform the benchmark index. In a similar vein, Jiang et al., 2021 [34] found that the performance gap between less skilled managers, who herd on the decisions of other institutional investors, and skilled managers, who antiherd, is larger during periods in which opportunities for active managers are more valuable, such as booms. Interestingly, Chen et al. (2021) [41] found evidence that fund manager skills may vary according to the market conditions when the manager entered the labor market. Fund managers who begin their careers during recessions exhibit superior timing skills during recessions, but lack strong picking skills during booms.

The findings above are all consistent with the idea that fund managers learn from experience. However, we note that the same quantitative relationship between skill and skill improvements may result from a cognitive bias that has been well established in the behavioral finance literature: conservatism bias. Conservatism bias is the reluctance to adapt to new information and make corresponding changes, often resulting in underreaction to new information [52]. Fund managers, for example, exhibit slow reactions to news such as firm earnings announcements, continuing to rely on historical information rather that fully incorporating new information [53]. In the case of foreign investment decisions, investors may rely on other sources of information, such as forecasts from local analysts, which may mean it takes longer to process and incorporate new information [54].

The degree of conservatism bias can be inferred from the portfolios of fund managers [55], but it is important to consider market efficiency when discussing conservatism bias. In less efficient markets, where stock prices react more slowly to new information overall, managers with less cognitive bias or faster reactions to new information have more opportunities to generate higher performance [56]. Since conservatism bias is an underreaction to new information, we expect that conservatism bias may be revealed in a failure of fund managers to improve their stock-picking and market-timing abilities, particularly for fund managers of funds currently exhibiting high risk-adjusted excess returns relative to the market.

Returning to our measures of stock picking and timing, we consider how the process of learning, or the existence of conservatism bias, would be revealed in our measures of stock-picking and market-timing ability. Hypothesis 2 examines the influence of conservatism bias on fund managers’ stock-picking and timing ability by examining the relationship between the current level of stock-picking or market-timing skill to improvements in stock-picking and market-timing skills.

Hypothesis 2 (H2).

Fund managers with poor stock-picking or market-timing skills are more likely to learn those skills and improve their skills in the following period.

2.3. Stock-Picking and Market-Timing Skill and Job Security

Thus far, our hypotheses have posited that (1) funds managed by fund managers with higher stock-picking and timing ability do, in fact, outperform other funds; and (2) fund managers with poor stock-picking or market-timing skills are more likely to learn those skills and improve in the following period Our final set of hypotheses examines whether the fund managers themselves reap the benefits of higher stock-picking and market-timing ability through higher job security and the related question of whether replacing fund managers—which we note may be voluntary or forced—improves fund performance.

Performance relative to peers plays a significant role in the length of a mutual fund manager’s career, and data suggest that those performance differences can be explained by manager characteristics [57,58]. Among those characteristics, stock picking-ability and market-timing ability have been found to be significant factors that affect excess returns, which is often considered one of the fundamental determinants of comprehensive performance [59]. Thus, we expect stock-picking ability and market-timing ability to be significant in explaining the likelihood of a fund manager being replaced.

Overall, there seems to be an inverse relation between the probability of managerial replacement and fund performance, although the relationship between managerial replacement and subsequent fund performance has also been shown to depend critically upon the skill of the replaced fund manager [60,61]. Further complicating the relationship between replacement and subsequent fund performance is the fact that replacement of a fund manager may be either forced or voluntary and may come at a time when the fund itself is performing relatively well or poorly, both of which seem to influence the post-replacement performance of funds [62,63,64].

Building on the existing literature, we propose the following hypotheses to examine the relationship between stock-picking and market-timing abilities and fund manager job security:

Hypotheses 3 (H3a).

Fund managers with high stock-picking or market-timing ability are less likely to be replaced, thereby enjoying greater job security.

Hypotheses 3 (H3b).

Replacement of a fund manager or adding a new fund manager in addition to the current manager leads to an improvement in fund performance.

3. Data and Methodology

3.1. Data

The data for this study were collected from the China East Money database. We selected all equity investment mutual funds that were established before 2018. For funds with multiple fee classes, we only included the A class, since any fee class will hold the same positions for a given fund. The selection process yielded a sample comprising 198 funds, with data reported on a quarterly basis, spanning from 2018 to 2022. Notably, Chinese manufacturing firms represent the largest sector, boasting the highest number of listed firms among all 19 sectors in the Chinese financial market. Fund returns are net asset value returns, factoring in the deduction of any management fees. Given that our sample period is very recent and discontinued funds are extremely rare in the Chinese market, the sample is not subject to survivorship bias. We estimate the timing and picking abilities of the fund managers using our new proposed measures—adapted versions of Equations (1) and (2)—as described above. Detailed summary statistics for the sample are presented in

Table 1. Summary statistics, overall sample. Fund manager skills, Timing and Picking, are defined in Equations (1) and (2). Fund manager learning is a dummy variable that takes the value of 1 if stock-picking skill improves (Learning_Picking), market-timing skill improves (Learning_Timing), or both improve (Learning). Fund outcomes include the Return of the fund, measured by the change in net asset value relative to the average return of all funds in the sample, and Replacement, a dummy variable that takes the value of 1 if the fund manager is replaced or a new manager is added. Fund characteristics include Size, the natural logarithm of fund total net assets; Expenses, the fund expense ratio; Turnover, the portfolio turnover ratio; Flow, the net outflow of funds out of the fund (net redemptions or purchases); Load, the total fund load; and Momentum, the fund’s yearly risk-adjusted relative return, or Sharpe ratio. Turnover is winsorized at the 1% level.

Statistic	N	Mean	St. Dev.	Min	Pctl (25)	Pctl (75)	Max
Fund Manager Skill
Picking	3762	−5.535	130.881	−2508.84	−39.484	36.335	1016.76
Timing	3762	17.574	234.871	−1434.72	−105.817	142.489	1839.60
Fund Manager Learning
Learning_Picking	3564	0.503	0.5	0	0	1	1
Learning_Timing	3564	0.468	0.499	0	0	1	1
Learning	3564	0.21	0.407	0	0	0	1
Fund Outcomes
Return	3762	0	6.644	−26.907	−4.04	3.624	45.396
Replacement	3762	0.058	0.234	0	0	0	1
Fund Characteristics
Size	3762	6.112	1.532	0.165	5.204	7.147	10.455
Expenses	3762	1.376	0.282	0.5	1.5	1.5	1.5
Turnover	3726	280.803	209.150	18.644	125.029	383.358	1338.154
Flow	3762	−0.011	0.186	−1.940	−0.044	0.007	3.306
Load	3762	1.449	0.139	0.6	1.5	1.5	1.5
Momentum	3762	0.64	1.109	−2.066	−0.178	1.545	2.658

Note: Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period of 2018–2022.

.

From Table 1, we note that the mean values of both stock-picking skill and market-timing skill are positioned close to the midpoint between the 25th and 75th percentiles. Overall, both variables exhibit approximately normal distributions. Interestingly, stock-picking skill has a negative mean but a smaller standard deviation, indicating that most managers struggle with stock picking. On the other hand, market-timing skill has a positive mean but a larger standard deviation, suggesting that managers are relatively better at timing the market due to entering at opportune moments. Manager skill—either stock-picking skill or market-timing skill—improves about half the time, but both skills improve for only about one out of five managers.

The relative returns of the funds in the sample range are quite variable, ranging from −27% to over 45%. Fund manager replacement seems to be a relatively rare phenomenon: a fund manager is replaced or another manager added in less than 6% of the 3762 fund-quarter observations covered in this sample.

The average fund in the sample is CNY 1297 million (USD 178 million), but the standard deviation is very large, and the funds in the sample range from CNY 1 million (USD 137,000) to CNY 34 billion (USD 4 billion). The average expense ratio and load ratio in about 1.4% of the fund’s total net assets and average fund experience a small positive inflow of new money into the fund at the equivalent of 100 million shares on average (a negative outflow or −0.011 in the adjusted units of 10 billion shares).

3.2. Methodology

3.2.1. Skill and Fund Performance

We begin by testing the main claim of this study, that the new sector-based measures of fund manager ability, stock picking, and market timing lead to higher fund returns. Thus, our first empirical tests are in-sample tests of the robustness of the sector-based stock-picking and timing measures resulting in higher fund returns, as posited in Hypothesis 1.

Hypothesis 1 (H1).

Fund managers with higher stock-picking or market-timing skill ability outperform their peers in both absolute and risk-adjusted relative returns.

To assess the impact of the fund managers’ stock picking and timing scores on the funds’ returns in the subsequent quarter relative to the benchmark return, we estimate the following regression model for fund i at time t:

$$Retur{n}_{i, t+1}={\beta }_0+{\beta }_{1}Skil{l}_{i,t}+{\mathrm{B}}_2{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t+1}$$

(3)

where:

$Retur{n}_{i, t+1}$ is the fund’s return relative to the average fund return in the same quarter;

$Skil{l}_{i,t}$ denotes either of our new measures, stock picking, $Pickin{g}_t^j$, or market timing, $Timin{g}_t^j$;

$X_{i,t}$ is a vector of fund-specific control variables, including size (the natural logarithm of total net assets under management in millions of dollars, ln (TNA)), expenses (subscription fee and management fee in % per year, Expenses), turnover rate (in % per year, Turnover), the percentage flow of new funds (Flow), and load (the sum of front-end and back-end loads and additional one-time fees charged to customers, Load). The lagged Sharpe ratio, i.e., the fund’s risk-adjusted relative return in the previous quarter, is also included to control for Momentum. To mitigate the impact of outliers on our estimates, and to be consistent with other studies of mutual fund skills such as Kacperczyk et al. (2014) [48], we winsorize Turnover at the 1% level. See Table 1 above for more details on variable definitions and computation.

If our newly proposed measures of stock picking and market timing are robust, then regression analysis of Equation (3) should yield positive coefficient estimates on skill; the coefficient estimate of β₁ in Equation (3) will be positive and statistically significantly different from zero.

3.2.2. Learning

To examine how fund managers incorporate new information and learn stock-picking and market-timing skills, we introduce dummy variables that indicate an increase in the stock-picking score, an increase in the market-timing score, or an increase in both scores for the subsequent period. These dummy variables are then subjected to Logit regression analysis, including the stock-picking ability score and market-timing ability score in the previous quarter as independent variables of interest. In this way, we investigate the causal relationship between past ability and improvements in the fund managers’ stock-picking and market-timing abilities, which we hypothesized above would be negative. The specific equation estimated is presented below in Equation (4).

$$Learnin{g}_{i, t+1}={\beta }_0+{\beta }_{1}Skil{l}_{i,t-1}+{\mathrm{B}}_2{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t+1}$$

(4)

where:

$Learnin{g}_{i, t+1}$ is a binary variable; if the next period’s picking (timing) ability is higher than that of the current period, it equals 1. Other variables are defined as above.

Hypothesis 2 posits that learning makes fund managers with relatively low stock-picking or market-timing skill more likely to improve their ability in the following quarter.

Hypothesis 2 (H2).

Fund managers with poor stock-picking or market-timing skills are more likely to learn those skills and improve their skills in the following period.

Thus, hypothesis 2 posits that the estimated coefficient of stock-picking and market-timing skill in Equation (4) will be negative.

3.2.3. Stock-Picking and Market-Timing Skills and Job Security

When fund managers underperform, upper management may resort to replacing fund management to enhance fund performance. Therefore, our final set of hypotheses centers on the relationship between stock-picking and market-timing ability and fund manager job security.

Hypotheses 3 (H3a).

Fund managers with high stock-picking or market-timing ability are less likely to be replaced, thereby enjoying greater job security.

Hypotheses 3 (H3b).

Replacement of a fund manager or adding a new fund manager in addition to the current manager leads to an improvement in fund performance.

To investigate the factors influencing such changes, we employed a Logit model in regression (5). These regressions were aimed to test the drivers of fund manager replacement within funds.

$$Replacemen{t}_{i, t}={\beta }_0+{\beta }_{1}Skil{l}_{i,t}+{\mathrm{B}}_2{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t}$$

(5)

where:

$Replacemen{t}_{i, t+1}$ is a binary variable; it equals to 1 if there is any change in the fund’s managers. Other variables are defined as above.

If higher stock-picking and market-timing skill reduces the likelihood of a fund manager being replaced, then we would expect the coefficient estimate on those skill variables to be statistically significantly negative: The parameter estimate β₁ in Equation (5) would be less than zero.

In our baseline specification, the in-sample test of our new measures of stock-picking and market-timing ability, we examined whether the fund performance of funds managed by fund managers with higher stock picking and market timing skill is in fact better. Above, we tested whether fund managers with higher stock-picking and market-timing skill are less likely to be replaced. However, we note that the replacement of a fund manager may be voluntary or forced. So, a related question to both is whether funds at which the fund manager is replaced demonstrate higher returns in the quarter following the replacement.

Equation (6) is employed to examine whether replacing the manager leads to higher returns for the fund in the following quarter. Equation (6) assesses the potential impact of manager replacement on the fund’s performance, as measured by both absolute returns and risk-adjusted returns relative to other funds, to evaluate whether a change in fund management leads to higher returns.

$$Retur{n}_{i, t+1}={\beta }_0+{\beta }_{1}Replacemen{t}_{i,t}+{\mathrm{B}}_2{\mathrm{X}}_{\mathrm{i},\mathrm{t}}+{\epsilon }_{i,t+1}$$

(6)

where:

$Retur{n}_{i, t+1}$ and other variables are defined as above.

4. Results

4.1. Manager Skill and Fund Performance

Table 2. Fund manager skill predicts relative return of the mutual fund. The dependent variable is future performance of the mutual fund relative to the average fund performance, Return, in the following quarter. The independent variables are our new measures of skill—stock picking and market timing—as well as various controls for fund characteristics. Those include Size, the natural logarithm of fund total net assets; Expenses, the fund expense ratio; Turnover, the portfolio turnover ratio; Flow, the net outflow of funds out of the fund (net redemptions or purchases); Load, the total fund load; and Momentum, the fund’s yearly risk-adjusted relative return, or Sharpe ratio. Turnover is winsorized at the 1% level.

	Dependent Variable
	Relative Return
	(1)	(2)	(3)	(4)
Stock-Picking Skill	0.012 ***		0.012 ***
	(0.0004)		(0.0004)
Market-Timing Skill		0.005 ***		0.005 ***
		(0.001)		(0.001)
Size			−0.207 ***	−0.162 **
			(0.070)	(0.077)
Expenses			1.974 ***	2.676 ***
			(0.575)	(0.633)
Turnover			−0.001	−0.001 *
			(0.0005)	(0.001)
Flow			0.142	−0.225
			(0.521)	(0.574)
Load			1.629	1.593
			(1.163)	(1.281)
Momentum			0.350 ***	0.294 ***
			(0.089)	(0.099)
Constant	−0.213 **	0.027	−4.020 ***	−4.878 ***
	(0.098)	(0.108)	(1.192)	(1.312)
Observations	3762	3762	3726	3726
R2	0.183	0.010	0.199	0.028
Adjusted R2	0.183	0.009	0.198	0.026
Residual Std. Error	6.007 (df = 3760)	6.613 (df = 3760)	5.934 (df = 3718)	6.537 (df = 3718)
F Statistic	841.304 *** (df = 1; 3760)	36.308 *** (df = 1; 3760)	132.014 *** (df = 7; 3718)	15.405 *** (df = 7; 3718)

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels; standard errors are shown in parentheses. Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period of 2018–2022.

presents the results of a test of whether our new measure of stock-picking and market-timing skill leads to higher-equity mutual fund returns (Hypothesis 1). In columns (1) and (2) of Table 2, we note that in the absence of control variables, the stock-picking and market-timing skill variables exhibit positive and statistically significant coefficient estimates, indicating that funds managed by fund managers with higher stock-picking and market-timing skills tend to perform better than other funds in the following quarter.

Turning to columns (3) and (4), which correspond to empirical estimation of Equation (3), the coefficient estimates of both stock-picking and market-timing skill remain positive and statistically significant even after accounting for control variables. This indicates that higher scores in both timing and picking contribute to an increase in the fund’s above-average return in the next quarter. Another interesting result reported in Table 2, columns (3) and (4) is the positive and highly statistically significant coefficient estimate on one of the control variables, expenses. We interpret this as evidence that value added by highly skilled managers is associated with higher costs. This would be consistent with existing theoretical and empirical research on the relationship between managerial skills and compensation (see, for example, Berk and van Binsbergen, 2015 [28], and Berk and Stanton, 2007 [65]). Finally, we note the highly statistically significant positive coefficient estimate on momentum, as measured by a long-term (annual) lagged Sharpe ratio. Hendricks, Patel, and Zeckhauser (1993); Goetzmann and Ibbotson (1994); and Brown and Goetzmann (1995) found evidence of persistence in mutual fund performance over short-term horizons of one to three years and attributed it to common investment strategies [66,67,68], while Grinblatt and Titman (1992) and Busse and Irvine (2006) attributed it to manager skill [69,70]. Other researchers have presented convincing evidence that this momentum effect may be mostly driven by superior performance funds [71] or poorly performing “losers” [19]. The reasons behind this finding may be a direction for future research, but we note here that the momentum effect also appears to exist, at least in the short term, in the Chinese market as well.

Overall, the results reported in Table 2 demonstrate that the new sector-weighted measures of stock-picking and market-timing skill pass our in-sample prediction test and may serve as indicators of the skill of mutual fund managers in stock picking and market timing, leading to higher than average returns on the funds they manage.

Further, Figure 1 and Figure 2 show that picking and timing abilities contribute to the fund’s relative returns. To further show that the picking and timing abilities reflect performance differences, we ranked the performance in each quarter. We put all samples in each quarter into ten groups, with the top group having 18 observations and all others having 20 observations. Then, the mean of each group in each period was recorded, and

Table 3. Relative performance based on picking scores.

Top	G2	G3	G4	G5	G6	G7	G8	G9	Bottom
5.651	3.410	2.083	1.343	0.626	−0.239	−1.249	−2.366	−3.378	−5.316
Top–Bottom sample mean difference	Standard error	G2–G9 sample mean difference		Standard error		Top–G2 sample mean difference		Standard error
10.968 ***	1.096	6.789 ***		0.742		2.241 **		1.028

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels.

and

Table 4. Relative performance based on timing scores.

Top	G2	G3	G4	G5	G6	G7	G8	G9	Bottom
1.424	1.064	0.884	0.631	0.313	−0.114	−1.230	−0.484	−0.784	−1.562
Top–Bottom sample mean difference	Standard error	G2–G9 sample mean difference		Standard error		Top–G7 sample mean difference		Standard error
2.988 **	1.232	1.849 **		0.740		2.656 ***		0.986

Note: ***, **, and * denote the statistical significance at the 1%, 5%, and 10% levels.

below show the mean of the mean vectors of each group. Then, for both picking and timing, we showed that the sample mean difference between the top and bottom groups is significant. We also showed that the sample mean difference between group 2 and group 9 is also significant. The results show that the higher-ability groups outperform the lower-ability groups. Lastly, we attempted to understand how many groups away would make the mean performance statistically significantly different in terms of picking and timing abilities. The top group and 2nd group in picking show significantly different performance. When testing timing, the top group and the 7th group showed significant performance differences. Such results show that the picking ability is more “valuable” than the timing ability, and the fund firms should give more value to emphasis on the manager’s evaluation if the performance is from picking.

4.2. Learning

Table 5. Low-skill fund managers improve. The dependent variables are fund manager learning: dummy variables that take the value of 1 if stock-picking skill improves (Learning_Picking), market-timing skill improves (Learning_Timing), or both improve (Learning). Learning is a dummy variable future performance of the mutual fund relative to the average fund performance, Return, in the following quarter. The independent variables are our new measures of skill—stock picking and market timing—as well as various controls for fund characteristics. Those include Size, the natural logarithm of fund total net assets; Expenses, the fund expense ratio; Turnover, the portfolio turnover ratio; Flow, the net outflow of funds out of the fund (net redemptions or purchases); Load, the total fund load; and Momentum, the fund’s yearly risk-adjusted relative return, or Sharpe ratio. Turnover is winsorized at the 1% level.

	Dependent Variable
	Learning_Picking Improved Stock-Picking Skill		Learning_Timing Improved Market-Timing Skill		Learning Improved Stock-Picking and Market-Timing Skill
	(1)	(2)	(3)	(4)	(5)	(6)
Stock-Picking Skill	−0.010 ***	−0.010 ***			−0.006 ***	−0.006 ***
	(0.0004)	(0.0004)			(0.0003)	(0.0003)
Market-Timing Skill			−0.031 ***	−0.031 ***	−0.008 ***	−0.008 ***
			(0.001)	(0.001)	(0.001)	(0.001)
Size	0.044	0.073 **	0.007	−0.010	0.007	0.015
	(0.031)	(0.031)	(0.033)	(0.033)	(0.034)	(0.035)
Expenses	0.266	0.217	0.142	0.134	0.270	0.266
	(0.235)	(0.237)	(0.260)	(0.260)	(0.283)	(0.283)
Turnover	0.0001	0.00001	−0.0003	−0.0003	−0.00001	−0.00002
	(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)
Flow	−0.160	−0.049	0.415 *	0.396	−0.320	−0.307
	(0.251)	(0.258)	(0.246)	(0.247)	(0.275)	(0.276)
Load	0.263	0.417	−0.531	−0.574	−0.070	−0.028
	(0.497)	(0.503)	(0.510)	(0.510)	(0.576)	(0.578)
Momentum		−0.306 ***		0.124 ***		−0.062
		(0.040)		(0.039)		(0.044)
Constant	−0.859	−0.995 *	0.416	0.519	−1.998 ***	−2.058 ***
	(0.538)	(0.545)	(0.520)	(0.521)	(0.598)	(0.601)
Observations	3528	3528	3528	3528	3528	3528

Note: ***, **, and * denote the statistical significance at the 1%, 5% and 10% levels; standard errors are shown in parentheses. Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period 2018–2022.

presents the results of an empirical estimation of Equation (4), which examines how a fund manager’s skill in the previous period affects improvements in fund manager skill in the following period through a Logit regression of improvements in a fund manager’s stock-picking skill, market-timing skill, or both, including past skill and various control variables.

The results reported in columns (1)–(2) of Table 3, based on an empirical estimation of Equation (4), demonstrate that poor stock-picking skill in the previous quarter leads to a statistically significantly higher likelihood of an increase in fund manager stock-picking skill in the following quarter. Similarly, in columns (3)–(4), we can observe that poor market-timing skill in the previous quarter leads to a statistically significantly higher likelihood of an increase in market-timing skill in the following quarter.

What about the relationship between past skill level and improvements in both stock-picking ability and market-timing ability? In columns (5)–(6), we note that the coefficient estimates of the previous quarter’s stock picking skill and market timing skill are both statistically significantly negative, indicating that lower skill in either stock picking or market timing in the previous quarter also leads to a higher likelihood of the fund manager improving both stock-picking and market-timing ability in the subsequent quarter.

Table 6. Average marginal effects of fund manager learning.

	AME1	AME2	AME3	AME4	AME5	AME6
Stock-Picking Skill	−0.0024 ***	−0.0025 ***			−0.0007 ***	−0.0007 ***
	(0.0001)	(0.0001)			(0.0000)	(0.0000)
Market-Timing Skill			−0.0078 ***	−0.0077 ***	−0.001 ***	−0.001 ***
			(0.0003)	(0.0003)	(0.0001)	(0.0001)
Size	0.011	0.0182 **	0.0017	−0.0026	0.0009	0.0019
	(0.0076)	(0.0078)	(0.0082)	(0.0083)	(0.0043)	(0.0044)
Expenses	0.0664	0.0543	0.0356	0.0334	0.0341	0.0336
	(0.0586)	(0.0593)	(0.0651)	(0.065)	(0.0358)	(0.0357)
Turnover	0.0000	0.0000	−0.0001	−0.0001	0.0000	0.0000
	(0.0001)	(0.0001)	(0.0001)	(0.0001)	(0.0000)	(0.0000)
Flow	−0.0401	−0.0123	0.1037 *	0.099	−0.0404	−0.0388
	(0.0627)	(0.0645)	(0.0614)	(0.0617)	(0.0348)	(0.0348)
Load	0.0658	0.1041	−0.1327	−0.1434	−0.0089	−0.0036
	(0.0586)	(0.1257)	(0.1274)	(0.1274)	(0.0728)	(0.0729)
Momentum		−0.0765 ***		0.0309 **		−0.0078
		(0.0099)		(0.0097)		(0.0056)

Note: ***, **, and * denote the statistical significance at the 1%, 5% and 10% levels; standard errors are shown in parentheses. Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period of 2018–2022.

displays the average marginal effects derived from the Logit regression model presented in Table 5. Notably, in both the categories of stock picking and market timing, there is a substantial difference of over 200 units between the 75th percentile and the 25th percentile. These results suggest that fund managers who have lower skill in either stock picking or market timing demonstrate more learning, or a greater improvement in those skills, in the following quarter.

Taken together, the results from columns (1)~(6) provide support for Hypothesis 3. Fund managers with poor stock-picking skills or poor market-timing skills are more likely to improve in either skill. In stating the initial hypothesis and here, we interpret this as evidence of more learning by fund managers with lower skills. However, we note that another way to interpret the same results would be that fund managers with high stock-picking and market-timing skills are less likely to improve their skills in the following period. This interpretation of the results would be consistent with the existence of a well-known cognitive bias known as conservatism bias. Fund managers with high stock-picking or market-timing skill may be slow to incorporate new information due to conservatism bias, and therefore less likely to improve their skills. However, without more detailed analysis, we highlight these results as possible evidence of conservatism bias and suggest this as a possible direction for future research on the topic of fund manager skill and performance in our conclusions below.

It is also worth noting that fund managers may increase their trading activities and change their investment styles when they feel other managers are faring better or when they discover important market signals. However, such responsiveness of portfolio allocations due to new public information decreases the manager’s skill [72]. In emerging markets, the market signal could be short but profitable. Some managers focus on the local news and update their trading targets and prices based on the rumors [73,74]. Such quick strategy changes are bilateral swords. If the news is verified and reflected in the economy, the funds outperform other peers [75]. If the news is not verified, the excessive stock position change increase the trading costs and cause the funds to underperform compared with other funds [76].

4.3. Manager Skill and Job Security

Table 7. Skill increases fund manager job security. The dependent variable is Replacement, a dummy variable that takes the value of 1 if the fund manager is replaced or a new manager is added. The independent variables are our new measures of skill—stock picking and market timing—as well as various controls for fund characteristics. Those include Size, the natural logarithm of fund total net assets; Expenses, the fund expense ratio; Turnover, the portfolio turnover ratio; Flow, the net outflow of funds out of the fund (net redemptions or purchases); Load, the total fund load; and Momentum, the fund’s yearly risk-adjusted relative return, or Sharpe ratio. Turnover is winsorized at the 1% level.

	Dependent Variable
	Replacement
	(1)	(2)
Stock-Picking Skill	−0.035
	(0.031)
Market-Timing Skill		−0.101 **
		(0.046)
Size	−0.124 **	−0.117 **
	(0.049)	(0.050)
Expenses	−1.191 ***	−1.216 ***
	(0.324)	(0.324)
Turnover	0.001 **	0.001 **
	(0.0003)	(0.0003)
Flow	−1.171 **	−1.121 **
	(0.461)	(0.456)
Load	2.352 ***	2.324 ***
	(0.777)	(0.775)
Momentum	−0.079	−0.092
	(0.063)	(0.064)
Constant	−4.028 ***	−4.009 ***
	(0.887)	(0.885)
Observations	3726	3726

Note: ***, **, and * denote the statistical significance at the 1%, 5% and 10% levels; standard errors are shown in parentheses. Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period of 2018–2022.

presents the results related to Hypotheses 3a: the question of whether higher stock-picking and market-timing skill lead to higher job security for fund managers. The columns in Table 5 report the results of an empirical Logit estimation of Equation (5): the relationship between fund manager skill and the likelihood of fund manager replacement.

Hypotheses 3 (H3a).

Fund managers with high stock-picking or market-timing ability are less likely to be replaced, thereby enjoying greater job security.

Contrary to our stated hypothesis, we note in column (1) that higher stock-picking skill of a fund manager does not statistically significantly reduce the manager’s risk of being replaced: the parameter estimate on stock-picking skill is negative, as expected, but not statistically significantly different from zero. In column (2), however, as expected, higher market-timing skill does statistically significantly reduce the fund manager’s risk of being replaced: the parameter estimate on market-timing skill is negative, as expected, and highly statistically significant.

In

Table 8. Average marginal effects of manager skill on job security.

	AME	AME
Stock-Picking Skill	−0.0018
	(0.0016)
Market-Timing Skill		−0.0052 **
		(0.0023)
Size	−0.0063	−0.0059 **
	(0.0025)	(0.0179)
Expenses	−0.0607 ***	−0.0618 ***
	(0.0164)	(0.0163)
Turnover	0.0000	0.0000 **
	(0.0000)	(0.0000)
Flow	−0.0597 *	−0.057 **
	(0.0231)	(0.0228)
Load	0.1199 **	0.1182 ***
	(0.0393)	(0.0391)
Momentum	−0.004	−0.0047
	(0.0032)	(0.0032)

Note: ***, **, and * denote the statistical significance at the 1%, 5% and 10% levels; standard errors are shown in parentheses. Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period of 2018–2022.

, the average marginal effect at the mean is reported. Each quarter’s additional unit of market-timing skill reduces the risk of being replaced by 0.42%. Recall from the summary statistics presented above that the market-timing skill variable has been rescaled so that the first 25th percentile is roughly 1, and the 75th percentile is approximately negative 1. Thus, fund managers with above average market-timing skills are 1% less likely to be replaced than other managers. While acknowledging that a 1% marginal effect is quantitatively small, we note that these are quarterly data, so for some managers, the accumulation after a few years may become quantitatively significant.

Taken together, the findings may be interpreted as suggesting that mutual fund firms place a higher emphasis on the market-timing skills of fund managers than on the stock-picking skills of fund managers. If that is the case, one explanation may be that high stock-picking ability contributes less to bottom-line profitability of the fund compared to high market-timing ability, a question that may be explored in further research.

Our final set of results, reported in

Table 9. Replacing managers does not improve fund relative performance. The dependent variable is future performance of the mutual fund relative to average fund performance, Return, in the following quarter. The independent variables include Replacement, a dummy variable that takes the value of 1 if the fund manager is replaced or a new manager is added, and our new measures of skill—stock picking and market timing—as well as various controls for fund characteristics. Those include Size, the natural logarithm of fund total net assets; Expenses, the fund expense ratio; Turnover, the portfolio turnover ratio; Flow, the net outflow of funds out of the fund (net redemptions or purchases); Load, the total fund load; and Momentum, the fund’s yearly risk-adjusted relative return, or Sharpe ratio. Turnover is winsorized at the 1% level.

	Dependent Variable
	Relative Return
	(1)	(2)	(3)
Fund Manager Replacement	−0.020	0.109	0.449
	(0.463)	(0.462)	(0.510)
Size		−0.143 *	−0.148 *
		(0.077)	(0.077)
Expenses		2.676 ***	2.656 ***
		(0.637)	(0.637)
Turnover		−0.001 *	−0.001 *
		(0.001)	(0.001)
Flow		−0.204	−0.230
		(0.577)	(0.577)
Load		1.525	1.562
		(1.288)	(1.288)
Momentum		0.227 **	0.264 ***
		(0.099)	(0.102)
Fund Manager Replacement × Sharpe			−0.661
			(0.423)
Constant	0.001	−4.866 ***	−4.888 ***
	(0.112)	(1.318)	(1.317)
Observations	3762	3726	3726
R2	0.00000	0.020	0.021
Adjusted R2	−0.0003	0.018	0.018
Residual Std. Error	6.645 (df = 3760)	6.564 (df = 3718)	6.563 (df = 3717)
F Statistic	0.002 (df = 1; 3760)	10.817 *** (df = 7; 3718)	9.774 *** (df = 8; 3717)

Note: ***, **, and * denote the statistical significance at the 1%, 5% and 10% levels; standard errors are shown in parentheses. Data source is the China East Money Database. All data are quarterly and cover 198 Chinese equity mutual funds over the period of 2018–2022.

, presents the results from estimation of Equation (6), providing evidence on the short-term effect of replacing a fund manager on fund returns. The empirical estimation of Equation (6) was designed to test Hypothesis H3b:

Hypothesis H3 (H3b).

Replacement of a fund manager or adding a new fund manager in addition to the current manager leads to an improvement in fund performance.

Turning to Table 7, we note that the coefficient estimate on the main variable of interest, Replacement, is statistically insignificantly different from zero in column (1) and remains statistically insignificant in columns (2) and (3), even after various controls are included.

The results reported in Table 9 imply that replacing a fund manager or adding additional managers does not necessarily lead to an increase in fund performance relative to other funds in the following quarter. Fully understanding this result may require further research. We note again that the replacement of a fund manager may be either voluntary or forced, and this may affect how replacement influences fund returns. Also, it is possible that managers require a significant amount of time to adapt to their new roles; understand the fund’s operations, investments, and strategies; and realize higher returns for the fund. We leave these questions for further research and conclude here that, while the empirical results do not definitively refute Hypothesis 3b, they also do not support Hypothesis 3b.

The career risk and abilities could also be reflected in the manager’s past job experience and social activities. The managers’ skill and decision making are related to their past job experience [77]. Fund managers may work as market analysts and possess advanced knowledge in some special regions and fields. The Chinese financial market institutions have a specific choice of universities when they hire new recruits. The managers could communicate closely if they graduated from the same university, and the alumni effect could influence the stock picking and timing [78]. From our results, the stock-picking skill seems more “valuable”, but the timing skill seems highly required by the mutual fund firms and significantly affects the managers’ career risks.

4.4. Summary of Findings

Table 10. Summary of findings.

Hypotheses	Validation	Discussion
H1: Fund managers with higher stock-picking or market-timing skill ability outperform their peers in both absolute and risk-adjusted relative returns.	Supported	Fund managers with higher stock-picking and market-timing skill earn statistically significantly higher fund returns relative to other funds.
H2: Fund managers with poor stock-picking or market-timing skills are more likely to learn those skills and improve their skills in the following period.	Supported	Fund managers with lower stock-picking and market-timing skill are more likely to learn and improve their skills in the following period.
H3a: Fund managers with high stock-picking or market-timing ability are less likely to be replaced, thereby enjoying greater job security.	Partly Supported	Fund managers with higher market-timing skills enjoy more job security, but stock-picking ability does not statistically significantly impact job security.
H3b: Replacement of a fund manager or adding a new fund manager in addition to the current manager leads to an improvement in fund performance.	Rejected	Replacement of a fund manager does not lead to statistically significantly higher fund returns relative to other funds.

provides a comprehensive summary of the findings and indicates whether each hypothesis is supported or rejected based on the results obtained throughout the study. It serves as a concise overview of the research outcomes and their alignment with the proposed hypotheses.

5. Conclusions and Directions for Future Research

In this study, a new approach is proposed for estimating fund managers’ stock-picking and market-timing ability using sector weights when full historical position information is not disclosed. In-sample prediction results demonstrate the robustness of this method, with higher measures of stock-picking and market-timing skill positively influencing future fund performance relative to other funds.

The study also uncovers evidence that stock-picking and market-timing skills are not static talents once is born with, but rather skills that are learned over time. Fund managers with poorer stock-picking and market-timing skill are more likely to improve their stock-picking ability or market-timing ability in the future, which is evidence of learning. We note that this result is also consistent with a documented cognitive bias known as conservatism bias. The results might also be attributed to fund managers with higher stock-picking and market-timing skills being less likely to absorb and implement new information, resulting in a lower likelihood of improving their skills. Future research may pursue this line as well and work to distinguish what underlying forces lead to this result.

Finally, this study analyzes the relationship between fund managers’ stock-picking and market-timing skills and job security. Logit analysis reveals that fund managers with better market-timing skills are less likely to be replaced, but the effects of better stock-picking skills on the likelihood of being replaced are insignificant. This suggests that the market values market-timing skills more than stock-picking skills. Our final set of empirical analyses demonstrates that when fund managers are replaced (as they often are), the fund does not enjoy any significant short-term increases in returns.

In our discussion of the empirical results, we hinted at several possible directions for future research. In addition to those closely related topics, future research directions could explore the relationship between fund managers’ skill and environmental, social, and governance (ESG) investments. Given the growing importance of ESG considerations, firms are increasingly facing costs and premiums for engaging in non-sustainable behaviors. Currently, sustainability issues have a more direct impact on returns and are linked to longer-term economic growth. It has been observed that a manager’s ESG investment not only aligns with sustainability goals, but also enhances stock returns and reduces portfolio volatility. This area presents a promising avenue for future investigation [79,80].

Similarly to how investors reduce uncertainty by diversifying their investments or purchasing insurance [81], managers facing uncertainty—particularly career risks—are willing to pay a high “risk premium” to mitigate these risks. The utility maximization problem for fund managers varies between bear and bull markets due to differing career risk factors in these environments. Since managers are more likely to be replaced during bear markets, and our current results indicate that timing is a crucial factor in decision making, it is worth investigating whether managers alter their strategies more quickly when they underperform at the start of a bear market. Existing research indicates that managers’ timing is better during bear markets [48]. The higher “risk premium” may prompt managers to swiftly change their strategies if they can find a safer alternative for their careers [82]. This raises a question for investors: Is this strategic timing due to an anticipation of negative market conditions [83], or is it a move driven by self-interest?