SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. I. Background and Hypotheses
  4. II. Data and Variables Measures
  5. III. Sample Statistics
  6. IV. Multivariate Analysis
  7. References

We examine the correlation between organizational structure (public vs. private) and managerial turnover in a large sample of United States offered mutual funds. Consistent with the hypothesis that publicly traded and privately held firms have different incentive structures and, as such, should differ in their treatment of internal control mechanisms, we find that public sponsors are more sensitive to prior fund performance when making replacement decisions and experience smaller post turnover performance improvements. Additional testing suggests a greater likelihood of fund manager replacement when mutual funds are team managed and when fund boards are more independent.

The concentration of ownership and how that ownership can be transferred, is a long running topic of interest among financial researchers (Berle and Means, 1932). From a corporate governance perspective, the primary interest is whether the effectiveness of internal and external control mechanisms varies across organizational structures.1 Despite this interest, our understanding of control mechanisms and their effectiveness remains limited to publicly held firms primarily because market-based performance metrics are generally not available for privately held firms. This is a shortcoming given the number and importance of private firms in labor markets. Mutual fund manager replacements provide a rare opportunity to test the impact of organizational form on internal control mechanisms as all investment companies, whether publicly held or privately owned, are required to provide market and operational performance data on the funds they sponsor. Mutual funds also have the added benefit of lower endogeneity concerns since sponsor organizational type is determined before funds are created and managers are appointed. Further, since sponsors oversee multiple funds, it is unlikely that an individual fund's performance impacts sponsor ownership.

In this paper, we employ a sample of 2,953 US sponsored mutual funds from 99 investment companies from 1999 to 2007 to examine how organizational structure influences the decisions to reward and discipline managers. In doing so, we attempt to answer the following three questions: 1) does organizational structure matter in the labor market for mutual fund managers? If so, 2) which type of firm, public or private, tends to be more sensitive to performance when making replacement decisions? and 3) how does organizational structure impact post managerial turnover fund performance? After controlling for fund performance and size, we document a significantly higher likelihood (about 10%) of fund manager replacement when sponsor firms are publicly, rather than privately, held. We also present compelling evidence that this increase is caused by the sensitivity of public sponsors to prior performance. Further, we document that post turnover performance improvements are smaller in public than private sponsor funds. Overall, our results suggest that ownership structure plays a crucial role in the fund manager turnover decision.

Our research contributes to the literature in three important ways. First, to the best of our knowledge, this is the first empirical evidence that provides a link between organizational form and managerial turnover. This separation of organizational form is necessary and relevant due to the sheer size of the private market and differing agency costs. The extant managerial turnover literature is almost exclusively limited to replacements that occur in public firms, yet private firms control considerable resources in most economies. Secondly, understanding the correlation between organizational form and managerial turnover is important as lower level managerial performance (e.g., fund managers) critically relates to the profitability of both mutual funds and the companies that sponsor them. If organizational structure influences how fund sponsors react to fund manager under-or overperformance, then organizational structure is an important consideration for shareholders of mutual funds. Finally, our study is the first to link internal governance mechanisms, such as team managed versus single managed, and board characteristics with managerial separation.

The remainder of the paper is organized into five sections. Section I provides the background literature and introduces the main hypotheses. Section 2 discusses the data generation process as it relates to firm organization and manager replacement. Section III presents the descriptive statistics and the performance measures used. Section IV provides the multivariate results of the study, while Section V provides our conclusions.

I. Background and Hypotheses

  1. Top of page
  2. Abstract
  3. I. Background and Hypotheses
  4. II. Data and Variables Measures
  5. III. Sample Statistics
  6. IV. Multivariate Analysis
  7. References

A. Organizational Structure and Fund Manager Turnover

The ownership structure of publicly held sponsors tends to be more atomistic and suffers more from the effects of agency costs than does the ownership of privately held ones (Adams, Mansi, and Nishikawa, 2010). The characteristics of diffused equity ownership, such as mandatory disclosure requirements and financial analyst coverage, lead managers of publicly traded firms to be sensitive to performance (Porter, 1992; Froot, Perold, and Stein, 1992; Loderer and Waelchli, 2010). This enhanced sensitivity to performance influences corporate decision making. Mortal and Reisel (2011) report that European Union publicly traded firms invest at higher rates and are more sensitive to investment opportunities, evidence they use to conclude that publicly traded firms allocate capital more efficiently than do privately held firms. Alternatively, heavy scrutiny from investors can prevent future investment and growth by influencing management decision making (Rappaport, 1990). For example, Asker, Farre-Mensa, and Ljungqvist (2011) find that US publicly traded firms whose stock prices are more sensitive to recent performance react less to changes in growth opportunities than do similar privately held firms. From a financing perspective, Giannetti (2003) finds that the market, legal, and regulatory environments in which the firms operate have a greater influence in privately held (i.e., unlisted) firms.

Ownership structure affects other firm policies. For instance, market participants react to dividend omissions by dramatically lowering their estimations of firm value (Michaely, Thaler, and Womack, 1995). Michaely and Roberts (2012) contend that this sensitivity to dividend policy leads publicly traded firms to smooth dividends more than private firms where the immediate change in value is less visible. They also find dividends from private firms, where information asymmetry problems are relatively small, are more sensitive to investment needs. Brav et al. (2005) report survey evidence that managers of private firms view the consequences of dividend cuts and omissions as less severe than their public counterparts.

Alternatively, the operating strategies of private sponsor funds may be more complex than public ones and incorporate concerns about firm and founder reputation in addition to measures of performance (Anderson, Mansi, and Reeb, 2003). Ferris and Yan (2009) also report that private fund sponsors charge lower fees and are less likely to be involved in fund scandals that harm sponsors’ reputations. Overall, the literature suggests that privately held firms have different incentive structures than publicly owned firms and, as such, should differ in their treatment of internal control mechanisms. This leads to our first testable hypothesis:

  • Hypothesis 1: 
    Fund manager replacements are more likely in mutual funds sponsored by publicly traded investment companies.

B. Organizational Structure and Performance

Sirri and Tufano (1998) report a positive relationship between prior period fund performance and subsequent cash inflows. Since fund sponsor revenues are a function of fund size, both public and private sponsors alike have incentives to monitor fund manager performance. Khorana (1996) finds that prior poor performance is associated with an increased probability of fund manager turnover, suggesting sponsors react to protect mutual fund shareholder wealth and to maximize sponsor revenues. Volpin (2002) and Coffee (1999) argue that higher sensitivity of managerial turnover to performance is indicative of effective corporate governance as underperforming managers have less opportunity to erode shareholder wealth. Brown, Harlow, and Starks (1996) contend that fund managers who are likely to be replaced tend to increase portfolio risk. Khorana (1996) notes this increased volatility has implications for fund manager replacement decisions as sponsors require more evidence, and implicitly more time, to accurately assess fund manager quality. The greater informational issues associated with publicly traded sponsor firms provide incentives for sponsor managers to be more aggressive when maximizing current period cash flows in order to signal their managerial value to sponsor shareholders. Because of the differing incentive structures of private and public sponsors, we hypothesize that private sponsors evaluate fund managers more subjectively. If so, we anticipate a greater sensitivity to performance for managerial turnover in funds that are publicly sponsored. In addition, we expect publicly sponsored funds’ greater sensitivity to performance to result in smaller post turnover performance improvements. This leads to our second and third testable hypotheses:

  • Hypothesis 2: 
    Publicly traded sponsors are more sensitive to fund performance in the managerial replacement decision-making process.
  • Hypothesis 3: 
    Performance improvements following fund manager turnover are smaller in funds sponsored by publicly traded investment companies.

C. Fund Management Structure and Boards of Directors

Our analysis controls for the impact of fund management structure on turnover decisions. We expect team managed funds to have higher rates of replacement than single managed funds as sponsors commonly use management teams to mentor new managers prior to the planned departures of senior managers. In addition, sponsors may find it easier to terminate individual team members. Massa, Reuter, and Zitzewitz (2010) note that turnover costs are lower for team managed funds as single managed funds experience more reductions in net flows following replacements. Alternatively, teams are often used in large funds where it is hard for a single manager to evaluate numerous securities, thereby making it more difficult for sponsors to identify underperforming managers. Similarly, management teams are routinely used in funds with complex investment objectives. Almazan et al. (2004) argue that fund manager career concerns are more compelling in individually managed funds since their reputation will be more affected by the successes or failures of these funds. They also argue that labor market discipline is less effective in team managed funds where it is difficult to identify individual contributions to fund performance (i.e., free rider problem). If turnover costs are lower in team managed funds and public sponsors are more sensitive to fund performance, we anticipate more frequent fund manager replacements in publicly sponsored team managed funds.

We also control for the influence of mutual fund boards of directors in the labor market for managers. While both have legal mandates to monitor management, mutual fund boards differ from corporate boards in that they have no employees, often include powerful insiders with opposing fiduciary responsibilities (i.e., officers and directors of the fund sponsor), and usually oversee several funds within the complex. Adams et al. (2010) document a significant correlation between board structure and mutual fund performance in public, but not in private offered funds. While the decision to hire, fire, and reward fund managers is made by sponsors, boards of mutual funds may affect the process.2 On the firm level, Weisbach (1988) posits that as boards become increasingly independent of managers, their monitoring effectiveness increases thereby decreasing managerial opportunism and enhancing overall firm performance.3

II. Data and Variables Measures

  1. Top of page
  2. Abstract
  3. I. Background and Hypotheses
  4. II. Data and Variables Measures
  5. III. Sample Statistics
  6. IV. Multivariate Analysis
  7. References

A. Sample

We utilize two primary databases in our analysis: 1) the Center for Research in Security Prices (CRSP) Survivor Free Mutual Fund database and 2) the Morningstar database. Both databases contain monthly class level returns and yearly information including total net assets (TNA), portfolio turnover, 12b-1 fees, expense ratios, age, purchase constraints such as institutional share classes, and historical returns. We rank the investment management companies listed in CRSP Mutual Fund database from largest to smallest in the year 2002 (mid-year of the sample) according to the size of the fund assets under management and select the 55 largest fund families. We then obtain data for small and midsize investment companies. We eliminate funds that are less than four years old since incubation policies bias the performance measures of relatively new funds where only successful incubated funds are offered for sale (Evans, 2010). We also eliminate fund families that do not employ managers to oversee actively managed funds (e.g., fund families that offer only index funds or exchange traded funds). The resulting database contains 99 investment management companies and represents over 80% of all mutual fund investments in 2002.

We determine whether the investment companies are publicly traded or privately held using the CRSP stock database. We also refer to Dun and Bradstreet's Million Dollar Directory and Hoover's Online to determine whether an investment company is a subsidiary of a publicly traded firm. For funds not listed in the aforementioned databases, we screen their website to determine ownership status or, in the case of subsidiaries, their parent's ownership status.4 We define public ownership as a dummy variable that is equal to one if the investment company is publicly traded and zero otherwise. For each fund, we also gather sponsor level information including the number of funds offered and the TNAs under management.

The Morningstar and CRSP databases list information on a per share class basis. Since the majority of mutual funds have multiple share classes that differ primarily in expenses, loads, and clientele, we combine the different classes into a single fund. Specifically, we compute fund level data for a given fund using the summation of the weighted class level data items, with the weight being the TNAs of each share class divided by total assets for the entire fund.

B. Measuring Fund Manager Turnover

We identify fund manager turnover by noting those instances in the Morningstar Principia database where the current fund manager was not the fund manager of record in the preceding year. ManagerChangei,t is a dummy variable that is equal to one if fund i experiences managerial turnover in year t. For completeness, we also verify fund manager replacement using the CRSP mutual fund database.5 It is common practice in the mutual fund industry to use teams to manage an individual fund or a group of funds (e.g., in 2002, approximately 57% of all funds were team managed). As such, we include a dummy variable that is equal to one if the funds are team managed (Team Managed). We identify fund manager replacement in team managed funds based on the start date of the longest tenured manager as senior managers generally control investment decisions (Qiu, 2003). We include team managed funds with anonymous managers only when Morningstar reports manager tenure information (about 4% of the sample). Finally, we exclude observations where funds have prior turnover event(s) in the preceding three years to insure prior performance measures are attributable only to the departing manager.

Manager replacement occurs for a variety of reasons including voluntary resignations, forced removals, retirements, promotions, and reassignments to other positions within the fund family. As a result, fund manager replacements can be associated with under, over, or average performance (Khorana, 1996). Since fund sponsors do not publicly identify the nature of the replacement, we follow Khorana (2001) and categorize fund manager replacements as forced when any of the performance measures used is negative. In other words, our analysis using this proxy will not estimate the likelihood of forced turnovers per se, but rather the likelihood of turnover in underperforming funds. We also utilize a calendar year matched control sample of funds that did not experience fund manager replacement in the preceding three years.

C. Measuring Performance Variables

Sirri and Tufano (1998) report a positive relationship between performance and cash flows into mutual funds. Since fund sponsors are compensated with fees based on portfolio size, they have powerful incentives to monitor manager performance. When considering whether to replace a manager, Morck, Shleifer, and Vishny (1989) argue that firms compare their manager's performance against the performance of other firms within the same industry. Similarly, Parrino (1997) documents an increased likelihood of manager replacement when reliable benchmarks are available.

In this research, we employ four benchmarks to measure fund performance. First, we compute each fund's objective adjusted return using the difference between the mean annual expense adjusted return of each fund's investment objective grouping and the corresponding return of each fund. That is:

  • image(1)

where OARi,t is the investment objective adjusted return of fund i in year t, Ri,t is the return of fund i in year t, and Ro,t is the return of an equally weighted index of funds with the same investment objective o in year t. Next, since the objective adjusted return does not account for risk, we compute Jensen's (1967) alpha using the S&P 500 as the benchmark index. For completeness, we also compute each fund's Carhart (1997) four factor alpha. The specifications for the Jensen (1967) and Carhart (1997) models are:

  • image(2)
  • image(3)

where Rindex is the return on the S&P 500 index, Rf, is the 30-day T-bill return, SMBt (small minus big) is the size factor that captures the stock return performance of small firms relative to large firms, HMLt (high minus low) is the relative return of value and growth stocks, and PR1YRt is a momentum factor computed as the difference in returns of prior year high and low return portfolios. The two alpha measures are computed over two year periods using monthly data.

Finally, we use annual net mutual fund flows (AnnualNetFlows) as our fourth performance measure. Khorana (1996, 2001) and Chevalier and Ellison (1999) find an increase in the probability of the fund manager's replacement when funds experience slow growth in annual net inflows. We compute net flows as:

  • image(4)

where TNAi,t is the TNAs in the fund i at the end of month t and Ri,t is the return of fund i during month t. We then compound the monthly annual net flow measures over each year.

D. Measuring Board Structure Variables

For our board structure variables, we collect calendar year end data from the statement of additional information (SAI), which is included in each fund's prospectus (Form 485). For each of the collected trustee level data items, we compute board level values using the average of each trustee level variable. Board variables include size, independent directors, and independent chair. Board Size is computed as the log of the number of trustees serving on each fund's board. Independent Directors is measured as the proportion of independent or disinterested trustees to total directors, where director independence is determined in accordance with SEC (2004) regulations. Under these regulations, independence indicates that an outsider is not an employee, not an employee family member, not an employee or a 5% shareholder of a registered broker-dealer, and is not affiliated with any recent legal counsel to the fund. Independent Chair is determined using manual collection based on Form 485 to ascertain whether the board has a chairperson who is an independent director. Similarly, we note from Form 485 those boards of directors that oversee all of the funds within a fund family complex (Unitary Board).

E. Additional Control Variables

Additional control variables are related to fund manager characteristics, clientele factors, and fund and family asset size. Fund manager characteristics include manager tenure and portfolio turnover. Khorana (1996) notes that better performing managers are less likely to be terminated than poor ones, suggesting a positive link between tenure and performance. We measure manager tenure (Manager Tenure) as the current or, in the case of replacement, the outgoing fund manager tenure in years. We follow Yan and Zhang (2009) and compute portfolio turnover (Portfolio Turnover) as the lesser of either the total amount of new securities purchased or sold divided by the average annual net fund asset value. Carhart (1997) and Gaspar, Massa, and Matos (2005) compute portfolio turnover similarly, but use the total of purchases and sales. The advantage of our portfolio turnover measure is that it more closely measures fund manager portfolio decisions by minimizing the impact of investor cash flows (Yan and Zhang, 2009).

Since investors have different expectations and sensitivities to fund performance, we also include two clientele effects: 1) fund expense ratio and 2) institutional holdings. Under the premise that some investors are willing to pay more for reputable fund management, we incorporate each fund's prior year expense ratio in our analysis (Expense Ratio). Similarly, we include institutional shareholders, computed as the ratio of shares held by institutions in the preceding year scaled by total fund shares (Institutional Holdings), since they evaluate fund manager performance differently than smaller investors.

We compute the log of fund and family assets under management (Fund TNA and Family TNA) to capture any variations in size. Larger mutual funds presumably require greater management skill than smaller funds making replacements more costly in larger funds, while larger sponsors may have deeper managerial talent pools than smaller ones making it easier to recruit new managers. We compute the log of fund age to capture variations in fund performance arising from the introduction of new funds since managers of newer funds may be evaluated differently than managers of older funds. Some sponsors use mutual fund incubation to develop new fund offerings, whereby successful startup funds are eventually offered to investors and unsuccessful ones are eliminated. Evans (2010) finds that incubation biases performance due to the ex post selection of superior performing funds. Fund incubation has important implications for manager turnover since managers of successful startup funds are less likely to be associated with poor performance. We also include investment objectives (Investment Objective) and year fixed effects throughout our analysis. Table I lists the definitions for the variables used in the analysis, as well as their data sources.

Table I.  Variable Definitions and Primary Source  For fund with multiple share classes we compute the weighted average value (using TNA of each class), where the reported fund TNA is the sum of the TNA from all classes. All Morningstar data are cross- checked or recomputed with the CRSP Mutual Fund database.
Variable Data Source Explanation
Performance and Risk
 Objective Adjusted ReturnMorningstar/CRSPComputed using two years of monthly data. Difference between a fund's return and its benchmark index in percentage.
 Annual Net FlowMorningstar/CRSPAnnual percentage growth rate in assets under management not due to portfolio returns.
 Jensen's AlphaMorningstar/CRSPComputed using two years of monthly returns, one month Treasury bills, and monthly S&P 500 returns in percentage.
 Carhart 4 Factor AlphaMorningstar/CRSPComputed using two years of monthly data. See Carhart (1997) for details.
 Growth RateMorningstar/CRSPThe two year investment objective adjusted growth rate of total net assets.
Fund and Family Characteristics
 Fund and Family TNAMorningstar/CRSPLog of total net assets of fund or family.
 Expense RatioMorningstar/CRSPPercentage of assets used to pay for operating expenses, including 12b-1 fees, management and administrative fees, and other asset-based costs incurred except for sales charges.
 Front LoadMorningstar/CRSPSales charge at initial fund purchase not included in expense ratio in percentage.
 12B-1 FeesMorningstar/CRSPPromotional and advertising expense charge included in expense ratio in percentage.
 Institutional HoldingMorningstar/CRSPPercentage of institutional class holdings in fund.
 Portfolio TurnoverMorningstar/CRSPTrading activity/change in portfolio holdings computed as the lesser of sales or purchases divided by average monthly total net assets in percentage.
Governance Characteristics
 PublicCRSP/Hoovers/Dun & BradstreetIndicator variable if a sponsor is publicly traded or the subsidiary of a publicly traded firm.
 Board SizeForm 485Log of number of directors on fund board.
 Independent DirectorsForm 485Proportion of directors who are classified as outsiders (independent).
 Independent ChairForm 485Dummy equal to one if the board chair is an independent director.
 Manager TenureForm 485The length of time, in years, the fund manager has directed the fund.
 Team ManagedForm 485Dummy equal to one if the fund has more than one manager.

III. Sample Statistics

  1. Top of page
  2. Abstract
  3. I. Background and Hypotheses
  4. II. Data and Variables Measures
  5. III. Sample Statistics
  6. IV. Multivariate Analysis
  7. References

A. Family Level Data

Table II provides descriptive statistics for the sample at the fund family level. Panel A reports the distribution of investment management companies’ assets segmented by ownership form. Included are the mean, median, and standard deviation for TNAs under management, market share, growth rate, and the number of funds. Due to skewness in the total asset and market share variables, the results diverge for mean and median values. For example, private investment companies have about a 147% larger asset base than public firms ($92 billion vs. $37 billion) and similar percentage of market share (1.72% vs. 0.73%), on average. However, the reverse is true for median values, which exhibit TNAs and market share that are larger for public than private sponsors ($21 billion vs. $14 billion and 0.44% vs. 0.26%). The mean and median two-year growth rates in TNAs under management are greater for private than public sponsors. Public sponsors provide their clients more investment options, offering about 22% more funds, on average, than private sponsors (39 funds vs. 32 funds). The difference is even more noticeable when median values are compared (36 funds vs. 24 funds).

Table II.  Descriptive Statistics (Family Level Data)  This table provides data on the nature of public and private ownership of investment management companies during a nine-year period beginning in 1999. The data set is comprised of 99 investment management companies as of 2002 and represents 762 firm-year observations. Total net assets are the year-end sum of investor funds in all objective categories. Market share is the percentage of the United States mutual fund market held by each investment management company. Growth rate is the two-year growth rate in total net assets in percentage terms and the number of funds is the total number of mutual funds sponsored by the investment management companies. Panel A reports the distribution of public and private investment management companies’ assets and related measures. Panel B lists the percentage of public investment management companies overall and in each size decile. Panel B also reports the percentage of public investment management companies in the top and bottom 5% size groupings.
Panel A. Investment Management Company Characteristics
Variables Public Ownership Private Ownership
Mean Median Standard Deviation Mean Median Standard Deviation
 
Total Net Assets37,23421,28744,63892,13513,985221,928
Market Share (%)0.7300.4430.8421.7150.2553.826
Growth Rate (%)34.06314.73884.14938.16626.81476.521
Number of Funds39.38236.0031.73931.73924.00041.132
Observations656565343434
Panel B. Investment Management Companies by Size Deciles
Size Deciles Public (%) Private (%)
Overall Sample65.634.4
Smallest 140.759.3
 268.032.0
 363.636.4
 463.236.8
 567.632.4
 677.222.8
 780.329.7
 881.818.2
 981.318.7
Largest 1050.050.0
Bottom 5%28.661.4
Top 5%21.278.8

Panel B presents the distribution of public and private investment management companies by deciles for the sample. Overall, for the years 1999-2007, public investment companies represent about 66% of the sample, while private investment companies account for about 34%. Public sponsor funds are found in greater proportion than their overall representation in most of the larger deciles (5, 6, 7, 8, and 9), while private sponsor funds are found in greater proportions in the smaller deciles (1, 3, and 4). Panel B also presents the top and bottom 5% size groupings. The data indicates that publicly owned status is less prevalent with the very smallest and largest sponsors, while the opposite occurs for private funds. Overall, the descriptive statistics presented in Table II suggest that public and private investment management companies differ in terms of characteristics such as size, market share, and fund offerings.

B. Fund Level Data

Panel A of Table III provides descriptive statistics for public and private investment companies at the fund objective level. Included are the mean, median, and standard deviations of relevant fund measures (fund, performance, and board variables) segmented by ownership status (public and private) for the overall sample. Panel A also reports the results of tests for differences in the mean and median values of each variable. Private funds represent the largest asset base in the data set with a median value of about 66% more than public sponsor funds (the difference in both mean and median values is significant at the 1% level). On average, expense ratios, front load fees, and 12B-1 fees are significantly higher for public than private sponsor funds (at the 1% level).

Table III.  Descriptive statistics (Fund Level Data)  This table reports descriptive statistics for the overall fund manager replacement and control sample that contains 10,429 mutual funds observations from 1999 to 2007. Panel A provides the mean, median, standard deviation, and number of observations for each variable and each ownership category. Mean and Median differences in the public and private values are also reported. Panel B presents the percentage of funds experiencing fund manager replacement by year for each ownership type. Panel C reports the percentage of funds experiencing fund manager replacement by investment objective category for each ownership type. Column 1 provides the investment object codes. Columns 2 and 3 represent the percentage of observations in the public and private samples, respectively. Columns 4 and 5 represent the percentage of fund manager replacements in public and private samples, respectively. Panel D presents correlation statistics for our sample. Variable definitions are provided in Table I.
Panel A. Public vs. Private Ownership
  Public Ownership Private Ownership Difference
Mean Median Standard Deviation Mean Median Standard Deviation Mean Median
 
Fund Characteristics
 Manager Tenure5.4985.0004.1216.0565.4004.656−0.558***−0.400***
 Number of Managers2.2292.0001.5212.0201.0001.4870.2091.000***
 TNA1,118.756336.5003,509.9433,407.446571.1009,861.227−2,288.690***−234.600***
 Fund Age14.91012.16710.84015.98013.84210.899−1.070***−1.675***
 Portfolio Turnover88.99753.000123.360114.73555.000243.488−25.73***−2.000
 Expense Ratio1.1201.0390.4611.0190.9800.5130.101***0.526***
 Front Load1.8501.6341.7241.3830.0001.7690.467***1.634***
 12B-1 Fees0.2590.2370.2420.1700.0170.2240.089***0.220***
  Growth Rate (%)12.901−6.90599.35726.9951.529104.914−14.094−8.434***
Performance Characteristics
 Objective Adjusted Return0.090−0.13011.1430.8120.36513.405−0.722***−0.495***
 Jensen's Alpha1.1050.8166.1071.5700.8587.115−0.465***−0.042**
 Carhart 4 Factor Model0.035−0.2015.1620.273−0.1196.206−0.238**−0.082
 Annual Net Flows54.239−3.3881,076.20421.7321.368239.06732.507*−4.756***
Board Characteristics
 Board Size8.8829.0002.6019.2749.0002.556−0.392***0.000***
 Independent Directors0.8090.8180.1210.7970.8000.0810.012***0.018***
 Independent Chair0.4340.0000.4960.4780.0000.500−0.044***0.000***
 Unitary Board0.3750.0000.4840.4420.0000.497−0.067***0.000***
 Observations   7,446  7,446   7,446   2,983  2,983   2,983  
Panel B. Incidence of Fund Manager Turnover by Sponsor Ownership Type
Manager Turnovers Public Ownership Private Ownership Single Managed Team Managed All Managed
Obs. (%) Obs. (%) Obs. (%) Obs. (%) Obs. (%)
 
19991795.87847.361198.911445.042636.27
20002317.571089.4616912.651705.953398.09
20012147.02726.30876.511996.972866.82
20022839.2813812.0818013.472418.4442110.04
200346515.251018.8413810.3342814.9956613.50
200433711.051018.841239.2131511.0343810.45
200551816.9816714.6217613.1750917.8268516.34
200646615.2822819.9718814.0750617.7269416.56
200735711.7014312.5215611.6834412.0550011.93
Total3,050100.001,142100.001,336100.002,856100.004,192100.00
Panel C. Fund Manager Replacements by Objective
Objective Title ICDI Code Public (%) Private (%) Public Replacements (%) Private Replacements (%) Single (%) Team (%) Single Replacements (%) Team Replacements (%)
  (1) (2) (3) (4) (5) (6) (7) (8) (9)
Aggressive GrowthAG8.619.6510.039.467.2710.448.4610.54
BalancedBL4.503.894.854.641.886.611.956.13
Global BondGB2.822.182.721.842.422.841.722.84
Global EquityGE2.542.922.724.121.703.531.573.82
Government SecurityGS3.695.403.975.433.674.664.274.41
IncomeIN1.452.181.802.191.601.711.352.17
International EquityIE8.048.889.647.277.758.799.818.61
Ginnie Mae FundGM2.191.412.160.881.842.081.571.93
Growth and IncomeGI6.187.887.546.485.777.516.217.74
High Quality BondBQ7.466.746.855.177.057.454.947.07
High Yield Money MarketMY1.090.540.920.531.050.820.750.84
High Quality MunicipalMQ6.804.435.343.947.115.185.844.55
High Yield BondBY2.782.983.152.192.772.902.623.01
Long Term GrowthLG13.3512.6413.9312.7012.9013.3812.9513.90
Precious MetalsPM0.401.210.490.440.970.320.670.39
Sector FundSF4.3510.594.6216.207.954.4414.604.59
Special/UnclassifiedSP0.050.000.030.000.020.060.070.00
Single State MunicipalMS19.5411.8014.5612.2622.9012.1017.2912.36
Total ReturnTR3.344.123.613.592.544.532.024.34
Utility FundsUT0.810.571.050.700.830.651.350.77
Observations 7,4462,9833,0501,1425,0475,3821,3362,856
(%) 100100100100100100100100
Panel D. Selected Pearson Correlations
  Objective Adjusted Return Annual Net Flows Public Ownership Team Managed Manager Tenure Family TNA Fund TNA Board Size Independent Directors
  1. ***Significant at the 0.01 level.

  2.  **Significant at the 0.05 level.

  3.   *Significant at the 0.10 level.

Annual Net Flows0.026***        
Public Ownership−0.028***0.016       
Team Managed−0.021***0.0130.086***      
Manager Tenure0.066***−0.012−0.060***−0.179***     
Family TNA0.053***−0.012−0.467***−0.154***0.032***    
Fund TNA0.046***−0.006−0.169***0.0190.181***0.327***   
Board Size0.0150.013−0.068***0.007−0.106***0.285***0.090***  
Independent−0.0070.0030.047***0.015−0.101***0.022***−0.0260.004 
Indep. Chair−0.0060.009−0.043***−0.012−0.128***0.201***0.0230.230***0.362***

Gaspar et al. (2005) contend that higher portfolio turnover suggests that mutual fund managers have shorter investment horizons. Alternatively, Ding and Wermers (2009) assert higher quality managers trade more often to exploit their superior security selection and market timing skills. Funds offered by privately held sponsors exhibit higher mean portfolio turnover values than the funds of publicly traded sponsors (115% vs. 89%). However, the two fund sponsor types have similar median values of portfolio turnover. Mean and median manager tenure are about half a year less in public versus privately sponsored funds with differences significant at the 1% level.6 The number of managers overseeing each fund is similar in both private and publicly sponsored funds.

In terms of performance, private funds have higher average objective adjusted returns, Jensen's (1967) alphas, and Carhart (1997) four factor alphas than publicly sponsored funds. The mean (median) annual net flows of publicly sponsored funds are significantly higher (lower) than those of privately sponsored funds. For board characteristics, board size and independence are similar in both public and private sponsor funds with median values of nine directors and independence ratios of about 80% (although the differences are statistically significant). On average, the independent chair ratio is slightly different in both categories and ranges from about 43% to 48% for public versus private ownership. Although not reported, in our sample, over 99% of directors oversee multiple funds within the fund family. Instead, we report the proportion of funds that utilize a single board (Unitary Board) to monitor all of the funds offered by their sponsor, which is greater for private than publicly sponsored funds (44% vs. 38%).

Panel B of Table III reports the incidence and frequency of fund manager replacement by year for the data used in the analysis for public, private, single managed, team managed, and all mutual funds. Panel B illustrates considerable variability in the annual incidence of fund manager replacement. Possible explanations for this variability include the growth in the number of funds available for sale, the mutual fund scandals of 2002-2003 documented in Ferris and Yan (2009), increased competition for managerial talent from hedge funds (Deuskar et al., 2011), and increased utilization of team management structure where replacement costs are likely lower. Fund manager replacement occurs in 3,050 (73%) public funds and 1,142 (27%) private funds during the sample period from 1999 to 2007 and is evenly distributed across both samples with most of the replacements occurring in the years 2003, 2005, and 2006. In terms of fund management structure, turnover occurs in 1,336 (32%) single managed and 2,856 (68%) team managed funds. The data suggest higher percentages of turnover in the earlier years of the sample for single managed funds (1999-2002) and more frequent percentage turnovers in team managed funds toward the later years of the sample (2003-2007).

Panel C provides data on replacements by fund objective. Public replacements have slight variations across each investment objective category, with the percentage of replacements in each category about the same as the representation of each category in the overall sample. Private replacements, however, occur less often in the growth and income category and more often in sector funds when compared to each category's representation in the overall sample. Team managed funds experience more replacements than single managed in funds with more complex investment strategies such as balanced, global bond, global equity, growth and income, and total return funds. However, single managed funds exhibit more turnovers than team managed funds in state municipal and sector funds. It is worth noting that public and private sponsors offer funds with similar investment objectives, so it is unlikely that differences in manager turnover rates are the result of investment objective preferences of a particular ownership structure.

Panel D reports the correlations for our key variables of interest. Fund performance, as measured by the two-year investment objective adjusted return, is negatively correlated with public ownership and team managed funds, and positively related to annual net flows, fund manager tenure, and fund and family size. We also find a negative and significant relationship between public ownership status and family TNAs. Overall, the descriptive statistics presented in Table III suggest that fund level characteristics vary across fund family ownership structure and investment objective.

IV. Multivariate Analysis

  1. Top of page
  2. Abstract
  3. I. Background and Hypotheses
  4. II. Data and Variables Measures
  5. III. Sample Statistics
  6. IV. Multivariate Analysis
  7. References

A. Logistic Analysis of Fund Manager Replacement

We provide multivariate logistic analysis to examine the association between ownership structure and fund manager replacement while controlling for fund characteristics. To test the hypothesis that sponsor level ownership structure is related to the incidence of fund manager replacements, we apply the following logit model and compute fund clustered errors as in Petersen (2009).7 That is:

  • image(5)

where ownership structure (PublicOwnership) and fund manager turnover (ManagerChange) are the primary variables of interest in the analysis. We estimate odds ratios to make the estimates easier to interpret. A value in excess of one on ownership type indicates a higher probability of turnover with public ownership and a lower probability of turnover with private ownership.

Table IV provides results for four model specifications. Model 1 uses each fund's prior two-year objective adjusted return to capture fund performance, Model 2 incorporates Jensen's (1967) alpha, Model 3 uses Carhart's (1997) four factor alpha, and Model 4 applies each fund's annual net flow measure.8 Overall, the results indicate a positive and significant relationship (at the 5% level) between public ownership and the likelihood of fund manager replacement. Across all models, the odds ratios range from about 1.16 to 1.18 indicating that the likelihood of turnover is about 10% higher in public versus privately sponsored funds. The remaining control variables that are applicable to all of the models have their expected signs. Consistent with Khorana (1996), prior performance using the investment objective adjusted return and Jensen's (1967) alpha (Models 1 and 2, respectively) is negatively and significantly related to fund manager replacement.

Table IV.  Fund Sponsor Ownership and Fund Manager Replacement  This table presents Logit regressions modeling impact of investment management company ownership type on the likelihood of fund manager replacement while controlling for board, clientele, and manager specific factors, fund size, and management company size. The data covers the period from 1999 to 2007 for 2,953 funds. Independent variables include a dummy variable that takes a value of one if the investment management company is publicly held (Public Ownership) using objective adjusted return (Model 1), Jensen's (1967) alpha (Model 2), Carhart's four Factor alpha (Model 3), Annual Net Flows (Model 4), the presence of multiple fund managers (Team Managed), the log of fund total net assets (Fund TNA), the log of fund age in years (Fund Age), the log of investment company total net assets (Family TNA), the log of the number of directors on the fund board (Board Size), the proportion of outside directors (Independent Directors), a dummy variable that is equal to one if the board chair is an independent director (Independent Chair Dummy) or if a single board oversees all family funds (Unitary Board), fund manager tenure in years (Manager Tenure), the approximate percentage of fund holdings that changed over the preceding year (Portfolio Turnover), the average expense ratio in the preceding year (Expense Ratio), and the proportion of equity holdings by institutions (Institutional Ownership). All models include year and investment objective fixed effects and are lagged one period. p-Values derived from fund-level clustered robust standard errors are in parentheses.
  Objective Adj. Return Jensen's Alpha Carhart Four Factor Alpha Net Flows
  (1) (2) (3) (4)
Public Ownershipt–11.1661.1721.1821.168
 (0.02)(0.02)(0.01)(0.02)
Team Managedt–11.5951.5931.6091.587
 (0.00)(0.00)(0.00)(0.00)
Objective Adjusted Returnt–2,t–10.988   
 (0.00)   
Jensen's Alphat–2,t–1 0.795  
  (0.00)  
Carhart 4 Factor Alphat–2,t–1  0.848 
   (0.26) 
Net Flowst–2,t–1   0.999
    (0.33)
Fund TNAt–10.9170.9210.9200.909
 (0.00)(0.00)(0.00)(0.00)
Fund Aget–11.3791.3831.3751.413
 (0.00)(0.00)(0.00)(0.00)
Family TNAt–11.0521.0481.0431.059
 (0.05)(0.07)(0.11)(0.03)
Board Sizet–11.0861.0881.1041.081
 (0.41)(0.84)(0.33)(0.45)
Independent Directorst–12.2702.2812.2612.238
 (0.00)(0.00)(0.00)(0.00)
Independent Chair Dummyt–11.0781.0791.0701.0583
 (0.23)(0.23)(0.23)(0.42)
Unitary Board Structuret–11.0911.0951.0871.125
 (0.16)(0.14)(0.18)(0.06)
Manager Tenuret–10.9350.9340.9340.934
 (0.00)(0.00)(0.00)(0.00)
Portfolio Turnovert–10.9660.9680.9760.973
 (0.01)(0.02)(0.08)(0.06)
Expense Ratiot–11.1261.1201.1251.134
 (0.12)(0.14)(0.13)(0.12)
Institutional Ownershipt–10.9780.9680.9700.991
 (0.80)(0.71)(0.73)(0.92)
Pseudo-R212.7012.5912.469.80
Log Likelihood−6,114−6,122−6,138−5,859
Model p-value0.000.000.000.00
Observations10,02210,02210,0229,729

The odds ratio estimate on the team managed dummy is greater than one and significant at the 1% level in all four models, reflecting the common industry practice of pairing new fund managers with more experienced (and often soon departing) managers.9 Another potential explanation is that replacement costs in team managed funds are lower than in single managed funds. In terms of board structure, the results for board independence are positively (e.g., odds ratio greater than one) and significantly related to fund manager turnover (at the 1% level). This is consistent with the idea that more independent boards are associated with better monitoring (Adams et al., 2010; Anderson, Mansi, and Reeb, 2004).

Fund size is negatively related to the likelihood of fund manager replacement, possibly reflecting the difficulty in finding replacements with the ability to manage large portfolios or that sponsors often promote superior managers to larger funds. Family TNA, however, is positively related to fund manager replacement (statistically significant in Models 1, 2, and 4 only) suggesting that larger fund families have a larger pool of potential managers than their smaller competitors. We also examine whether the increased likelihood of fund manager replacement by publicly traded sponsors is because they offer more funds and, as a result, have larger candidate pools. Although not reported, we find that public sponsors are significantly more likely to replace fund managers after controlling for the number of funds offered by sponsors. Fund age is positively and significantly related to manager turnover in all four models, consistent with the idea that managers of newer funds are evaluated differently than managers of older ones. The positive and significant estimated coefficient on fund age may reflect incubation policies where only superior performing funds (i.e., presumably those with superior managers who face lower threats of dismissal) are eventually offered to investors (Evans, 2010).

Similar to Chevalier and Ellison (1999), the manager tenure estimated coefficient is negative and significant suggesting that longer tenured managers are less likely to be replaced. Unlike Khorana (2001), we report that portfolio turnover is negatively and significantly related to the likelihood of subsequent period replacement implying that managers with more active trading strategies are less likely to be replaced since they have better private information and/or greater stock picking and market timing abilities (Ding and Wermers, 2009).

The results in Table IV demonstrate an increased likelihood of fund manager replacement when the shares of the sponsor are publicly rather than privately traded, experience poor performance, are team managed, and have independent boards. The evidence indicates that sponsor level attributes play an important role in governance in general and in fund manager replacement in particular.

B. Sponsor Ownership, Fund Management Structure, and Replacement

Table V provides subsample analysis for the impact of fund management structure (single vs. team managed) and sponsor ownership type on fund manager replacement decisions. Table IV reports an increased likelihood of fund manager replacement when funds are team managed suggesting lower turnover costs in team managed funds. In addition, our results are consistent with the common industry practice of pairing newer managers with senior ones who may be more likely to retire or otherwise depart. Management teams are also employed in more complex and diversified funds. If public sponsors are more sensitive to fund performance, we expect public ownership status to have a greater impact on team managed funds where turnover costs are lower. Similarly, the impact of fund management structure on the likelihood of replacements may vary in public versus private sponsors.

Table V.  Sponsor Ownership and Team Management Structure: Subsample Logit Analysis  This table presents Logit regressions modeling the impact of investment management company ownership type on the likelihood of fund manager replacement while controlling for board, clientele and manager specific factors, fund size, and management company size. The data covers the period from 1999 to 2007 for 2,953 funds. Independent variables include a dummy variable that takes a value of one if the investment management company is publicly held (Public Ownership), objective adjusted return for two years prior to the replacement (Objective Adjusted Return), growth in fund assets not due to portfolio returns for two years prior to replacement (Net Flows), the presence of multiple fund managers (Team Managed), the log of fund total net assets (Fund TNA), the log of the fund age in years (Fund Age), the log of the investment company total net assets (Family TNA), the log of the number of directors on the fund board (Board Size), the proportion of outside directors (Independent Directors), a dummy variable that is equal to one if the board chair is an independent director (Independent Chair Dummy) or if a single board oversees all family funds (Unitary Board), fund manager tenure in years (Manager Tenure), the approximate percentage of fund holdings that changed over the preceding year (Portfolio Turnover), the average expense ratio in the preceding year (Expense Ratio), and the proportion of equity holdings by institutions (Institutional Ownership). All models include year and investment objective fixed effects and are lagged one period. p-Values derived from fund-level clustered robust standard errors are in parentheses.
  Public Funds Private Funds Single Managed Team Managed Interaction (1) × (4)
  (1) (2) (3) (4) (5)
Public Ownershipt–1  1.0021.4321.007
   (0.98)(0.00)(0.93)
Team Managedt–11.7541.417  1.301
 (0.00)(0.00)  (0.01)
Public Own.*Team Managed    1.318
     (0.02)
Objective Adjusted Returnt–2,t–10.9840.9930.9880.9860.989
 (0.00)(0.01)(0.00)(0.00)(0.00)
Net Flowst–2,t–10.9990.9991.0020.9990.999
 (0.84)(0.32)(0.10)(0.34)(0.32)
Fund TNA0.9440.8970.8870.9760.917
 (0.05)(0.01)(0.00)(0.49)(0.00)
Fund Age0.9441.2631.1561.5141.396
 (0.00)(0.02)(0.06)(0.00)(0.00)
Family TNA0.9411.1180.9001.0981.053
 (0.12)(0.01)(0.01)(0.03)(0.05)
Board Size1.0241.0341.6680.8111.058
 (0.84)(0.89)(0.00)(0.16)(0.58)
Independent Directors3.0701.3343.2992.2682.389
 (0.00)(0.66)(0.00)(0.03)(0.00)
Independent Chair Dummy0.9141.1981.1970.9961.038
 (0.24)(0.18)(0.05)(0.97)(0.56)
Unitary Board Structure1.1501.1000.9071.0101.118
 (0.07)(0.43)(0.27)(0.91)(0.08)
Manager Tenure0.9420.9390.9030.9640.936
 (0.00)(0.00)(0.00)(0.01)(0.00)
Portfolio Turnovert–11.0390.9220.9400.9990.965
 (0.21)(0.00)(0.00)(0.95)(0.01)
Expense Ratiot–11.2141.0611.1181.1461.131
 (0.05)(0.69)(0.29)(1.21)(0.12)
Institutional Ownership0.9321.1371.0940.9220.999
 (0.51)(0.51)(0.48)(0.95)(0.99)
Pseudo-R210.5113.8514.737.6710.21
Log Likelihood−4,108−1,638−2,827−2,878−5,832
Model p-value0.000.000.000.000.00
Observations6,8432,8865,1964,5339,729

Models 1 and 2 provide results for public and private funds, while Models 3 and 4 provide segmentations for single and team managed funds. Model 5 combines all specifications and adds an interaction term to capture funds that have public ownership and are team managed. Models 1 and 2 report significant results on the team managed dummy although the economic significance is much greater for public sponsors where the odds of team managed replacement are about 1.75 times greater than the odds of replacement at single manager funds. For private sponsor funds, the odds of team managed replacement are about 1.42 times those of single managed funds. The odds ratios reported in Models 1 and 2 indicate that that the probability of turnover in team managed funds is about 23% greater in public versus privately sponsored funds. Similarly, Models 1 and 2 find that board independence is significant only for publicly traded sponsors. This finding is consistent with recent research by Adams et al. (2010), who find a positive impact of board monitoring on operational performance in publicly sponsored funds where agency costs are particularly high.

For funds with only one manager (Model 3), public ownership type is not significantly related to manager replacement. However, public sponsor ownership is associated with a statistically significant likelihood of replacement (at the 1% level) in team managed funds (Model 4). In Model 5, the estimated odds ratio for public sponsor ownership type is insignificant, while the odds ratios both for the team managed dummy and the public ownership-team managed interaction term are positive and significant (at the 1% and 5% levels, respectively). This finding suggests that publicly traded sponsors evaluate managers in team managed funds differently than privately held sponsors.

Similar to the results in Table IV, we find that prior fund performance is negatively and significantly related to replacement decisions in both fund management types (odds ratios less than one). Interestingly, portfolio turnover is positively and significantly related to fund manager replacements in private and single manager funds only. Additionally, the estimated coefficient on manager tenure is negative and significant for all four subsamples. Overall, the results presented in Table V suggest that sponsor ownership and fund management structures are important determinants of fund manager replacement decisions.

C. Turnover Sensitivity to Performance

Although the results in Table IV provide strong evidence that sponsor ownership type is significantly related to the likelihood of fund manager turnover, they do not tell us how these internal governance mechanisms impact the nature of the turnover event (e.g., forced vs. voluntary turnover decisions). Research suggests that prior performance is an important determinant of whether a manager is voluntarily replaced or forcibly removed (Weisbach, 1988; Bonnier and Bruner, 1989; Furtado and Rozeff, 1987). If agency considerations associated with dispersed ownership causes public sponsors to be more sensitive to performance than private sponsors, they should be more likely to terminate underperforming managers. Sponsor level cash flows are a function of the performance of individual mutual funds in the fund complex where more important (e.g., larger) funds impact sponsor level cash flows more than smaller, less important funds. If publicly traded investment companies are more sensitive to prior performance than privately held ones, then they should be especially sensitive to the performance of their largest, most important funds as they represent a greater impact on overall sponsor cash flows. To test these hypotheses, we interact prior performance with public sponsor ownership and team management and repeat the logit analysis of Table IV.

Table VI reports different interaction model specifications. Model 1 interacts prior fund performance with public ownership status. Model 2 repeats the interaction analysis using the team management dummy. Models 3 and 4 provide segmentations based on the relative size of each fund, where relative size is computed as the ratio of fund TNA to family TNA. Funds whose relative size is less than the sample median are included in Model 3, while funds with relative size measures greater than the sample median are included in Model 4.

Table VI.  Fund Sponsor Ownership and Performance Sensitivity  This table presents Logit regressions modeling the impact of investment management company ownership type on the likelihood of fund manager replacement while controlling for board, clientele, and manager specific factors, fund size, and management company size. The data covers the period from 1999 to 2007 for 2,953 funds. Independent variables include a dummy variable that takes a value of one if the investment management company is publicly held (Public Ownership), objective adjusted returns for two years prior to the replacement (Objective Adjusted Return), growth in fund assets not due to portfolio returns for two years prior to replacement (Net Flows), the presence of multiple fund managers (Team Managed), the log of the fund total net assets (Fund TNA), the log of the fund age in years (Fund Age), the log of the investment company total net assets (Family TNA), the log of the number of directors on the fund board (Board Size), the proportion of outside directors (Independent Directors), a dummy variable that is equal to one if the board chair is an independent director (Independent Chair Dummy) or if a single boards oversees all family funds (Unitary Board), fund manager tenure in years (Manager Tenure), the approximate percentage of fund holdings that changed over the preceding year (Portfolio Turnover), the average expense ratio in the preceding year (Expense Ratio), and the proportion of equity holdings by institutions (Institutional Ownership). All models include year and investment objective fixed effects and are lagged one period. p-values derived from fund-level clustered robust standard errors are in parentheses.
  Public Ownership Team Managed Less Important More Important
  (1) (2) (3) (4)
Public Ownershipt–11.1581.1540.9611.443
 (0.03)(0.04)(0.70)(0.00)
Objective Adjusted Returnt–2,t–10.9940.9890.9920.998
 (0.06)(0.00)(0.06)(0.67)
Net Flowst–2,t–10.9990.9990.9990.999
 (0.29)(0.27)(0.41)(0.13)
Team Managedt–11.5831.5861.6181.674
 (0.00)(0.00)(0.00)(0.00)
Fund TNAt–10.9140.9140.9430.924
 (0.00)(0.00)(0.15)(0.10)
Fund Aget–11.3911.3941.3721.373
 (0.00)(0.00)(0.00)(0.00)
Family TNAt–11.0561.0591.0191.010
 (0.04)(0.03)(0.68)(0.85)
Board Sizet–11.0811.0801.1930.996
 (0.76)(0.45)(0.23)(0.98)
Independent Directorst–12.3142.3041.8893.567
 (0.00)(0.00)(0.15)(0.00)
Independent Chair Dummyt–11.0501.0531.0870.953
 (0.45)(0.42)(0.37)(0.60)
Unitary Board Structuret–11.1171.1181.0281.151
 (0.08)(0.08)(0.76)(0.11)
Portfolio Turnovert–10.9680.9670.9311.031
 (0.02)(0.02)(0.00)(0.29)
Manager Tenuret–10.9360.9360.9310.951
 (0.00)(0.00)(0.00)(0.00)
Expense Ratiot–11.1341.1311.1240.995
 (0.12)(0.13)(0.05)(0.97)
Institutional Ownershipt–11.0031.0031.1320.865
 (0.98)(0.98)(0.38)(0.24)
Public*Obj. Adj. Returnt–2,t–10.989 0.9900.986
 (0.01) (0.11)(0.02)
Team*Obj. Adj. Returnt–2,t–1 0.998  
  (0.58)  
Pseudo-R210.2210.1611.9210.60
Log Likelihood−5,831−5,836−2,866−2,896
Model p-value0.000.000.000.00
Observations9,7299,7294,8454,884

The level and significance of the public sponsor dummy variable and the other governance related measures reported in Models 1 and 2 are consistent with the results presented in Table IV and support our hypothesis that public companies are more likely to terminate or promote fund managers. The significant odds ratio estimate on the interaction between public sponsor ownership type and each fund's objective adjusted return (Model 1) imply that public sponsors are more sensitive to prior performance in the fund manager replacement decision. The odds ratio is also economically significant indicating that the probability of turnover in publicly sponsored funds increases by an additional 4% relative to private funds when performance decreases by about 5% (our interquartile range). The odds ratio on the interaction between the team managed dummy and prior performance variables is less than one, though not statistically significant in Model 2.

For the subsample comprised of funds, that are relatively small from the sponsor's perspective, Model 3 indicates that the public ownership dummy variable is negatively (estimated odds ratio is less than one), although not significantly related to the likelihood of fund manager replacement. However, Model 4 reports that for those funds whose ratio of fund to family TNA is above the sample median (e.g., more important), public ownership structure is associated with an increased likelihood of fund manager replacement (results that are statistically significant at the 1% level). In terms of sensitivity to prior fund performance, Models 3 and 4 report that the interaction between public ownership type and prior fund performance (Public* Obj. Adj. Returnt–2,t–1) is statistically significant for relatively large funds (Model 4) only. In terms of economic significance, the public ownership results reported in Model 4 indicate that fund manager replacements in relatively large funds are about 26% more likely in public versus private funds when performance decreases by 5%.

The results presented in Table VI demonstrate that the estimated odds ratios for sponsor ownership structure are similar for overall and performance related turnovers. These findings suggest that the manager evaluation and replacement processes are different in public versus private firms. Public firms aggressively evaluate fund managers and have a higher sensitivity to prior performance.

D. Post Turnover Performance

In this section, we examine the consequences of fund manager replacement decisions. If public sponsors are more sensitive to performance than their private counterparts, and replace underperforming managers more frequently, underperforming managers will have less opportunity to erode fund value. Alternatively, private sponsors may evaluate fund managers subjectively and require greater levels of underperformance than public sponsors (Bushee, 2001; Porter, 1992; Shleifer and Vishny, 1989). If so, the expected ability of the new private sponsor fund manager should exceed the ability of the previous manager by a greater margin than in public sponsor replacements. Both hypotheses predict relatively small performance improvements in publicly sponsored funds following forced turnover events.

Table VII reports ordinary least squares estimated coefficients for changes in our primary performance measure, each fund's objective adjusted return. We deduct the value computed for the two years prior to the replacement calendar year end from the value computed over the two years after the replacement (e.g., t+1, t+2 vs. t–2, t–1).10 To control for the nature of the turnover event, we include a forced turnover dummy variable in each model that is equal to one if there is a turnover event and the objective adjusted return (computed over t–2 and t–1) is negative. We acknowledge that some replacements at poorly performing funds may be voluntary. To address concerns that misidentification of forced turnover events drives our results, we follow Campbell et al. (2011) and 1) employ a two stage Heckman approach and 2) use an alternate specification consisting of a voluntary turnover dummy and voluntary/public interaction terms to verify the coefficients are of opposite sign and significance. The results of these alternative specifications support the results presented in Table VII. In addition, we include variables to capture the interactions between the forced dummy and internal governance and fund management measures. Each model also includes the expense ratio, portfolio turnover, institutional ownership, fund and family size, and dummies for year and investment objectives.

Table VII.  Performance Changes Following Fund Manager Replacements  This table presents ordinary least square regressions of two-year changes in mutual fund performance on investment management company ownership type. Models 1-4 classify replacements as forced if the Objective Adjusted Returnt–2,t–1 is negative. The data covers the period from 1999 to 2007 for 2,953 funds. The dependent variables are the difference in the Objective Adjusted Returnt–2,t–1 in the two-year period following manager replacement and the prior two year period. Independent variables include a dummy variable that takes on a value of one if the turnover is forced (Forced Turnover), a dummy variable that takes a value of one if the investment management company is publicly held (Public Ownership), the presence of multiple fund managers (Team Managed), the log of the fund total net assets (Fund TNA), the log of the fund age in years (Fund Age), the log of the investment company total net assets (Family TNA), the log of the number of directors on the fund board (Board Size), the proportion of directors who are classified as outsiders (Independent Directors), a dummy variable that is equal to one if the board chair is an independent director (Independent Chair Dummy) or if a single board oversees all family funds (Unitary Board), fund manager tenure in years (Manager Tenure), the approximate percentage of fund holdings that changed over the preceding year (Portfolio Turnover), the average expense ratio in the preceding year (Expense Ratio), and the proportion of equity holdings by institutions (Institutional Ownership). All models include year and investment objective fixed effects. Fund-level clustered robust t-statistics are in parentheses.
  Public Ownership Team Managed Board Size Independent Directors Independent Chair
  (1) (2) (3) (4) (5)
  1. ***Significant at the 0.01 level.

  2.  **Significant at the 0.05 level.

  3.   *Significant at the 0.10 level.

Forced18.820***15.932***1.26829.026***14.880***
 (14.17)(14.95)(0.36)(8.02)(19.80)
Public Ownership0.781−0.453−0.535−0.433−0.502
 (1.40)(0.95)(1.12)(0.91)(1.05)
Team Managed−0.697*−0.179−0.772*−0.771*−0.796**
 (1.72)(0.40)(1.91)(1.92)(1.98)
Board Size−0.113−0.121−0.264***−0.127−0.132
 (1.34)(1.42)(2.72)(1.49)(1.55)
Independent Directors−2.031−2.404−2.4340.424−2.511
 (1.24)(1.47)(1.48)(0.24)(1.54)
Independent Chair Dummy−0.531−0.455−0.379−0.3000.103
 (1.25)(1.08)(0.90)(0.71)(0.21)
Unitary Board0.1900.1460.1120.1070.170
 (0.37)(0.29)(0.22)(0.21)(0.33)
Portfolio Turnover0.1660.1460.1200.1390.142
 (1.05)(0.92)(0.76)(0.88)(0.90)
Expense Ratio1.477**1.547**1.519**1.495**1.566**
 (2.23)(2.35)(2.30)(2.27)(2.38)
Institutional Ownership0.5810.6980.6690.6120.726
 (0.96)(1.16)(1.11)(1.01)(1.20)
Fund TNA−1.186***−1.174***−1.224***−1.208***−1.204***
 (6.15)(6.10)(6.27)(6.24)(6.21)
Fund Age1.160***1.114***1.131***1.178***1.131***
 (2.71)(2.60)(2.62)(2.74)(2.63)
Family TNA0.569***0.551***0.575***0.554***0.597***
 (2.70)(2.61)(2.72)(2.62)(2.81)
Public Ownership* Forced−6.717***    
 (4.80)    
Team Managed*Forced −3.270***   
  (2.84)   
Board Size* Forced  5.771***  
   (3.45)  
Indep. Directors* Forced   −18.995*** 
    (4.46) 
Independent Chair* Forced    −2.536**
     (2.46)
Unitary*Forced     
Adjusted R212.0011.6411.6811.7111.59
Model p-value0.000.000.000.000.00
Observations7,4187,4187,4187,4187,418

Model 1 of Table VII focuses on forced turnover decisions, sponsor ownership status, and post turnover performance. Model 1 reports that underperforming funds that experience fund manager turnover (forced) are associated with about a 19% improvement in performance, results that are statistically significant at the 1% level (similar to results reported in Khorana, 2001 on mutual fund manager replacements and in Denis and Denis, 1995 for CEO replacements).11 However, the forced and public sponsor interaction term is negative and significant at the 1% level indicating that performance improvements in publicly sponsored funds where the manager is terminated are about 7% lower than in private funds. This finding is consistent with the hypothesis that the fund manager replacement process and its consequences are different in private versus public firms. Model 2 examines the interaction between team managed funds and forced turnover. The estimated coefficient on the interaction term is negative and significant (at the 1% level). This suggests that while turnover is more likely in team managed funds, performance improvements following replacement events are about 3% higher in single managed funds.

Model 3 focuses on board size and replacements and reports a positive and statistically significant correlation between changes in fund performance and the board size/forced turnover interaction term. These findings suggest that turnover is more likely at funds with large boards, and that post turnover performance is also better. Model 4 reports that the proportion of independent directors/forced turnover interaction term is negative (about 19 basis points) and significant at the 1% level. Model 5 reports similar, albeit economically smaller, results for independent chairs. In unreported regressions, we find the unitary board/forced turnover interaction coefficient is insignificant.12 We also find that mutual fund boards are more independent following turnover, but this increased board independence is unrelated to post turnover performance improvements. The lower post replacement performance for more independent boards in Models 4 and 5 are consistent with independent boards being less tolerant of poor performance and more likely to petition sponsors to replace managers in order to prevent erosion of shareholder wealth.

V. Conclusion

This paper investigates whether the incentive structures of privately held and publicly owned firms impact their treatment of internal control mechanisms. In doing so, we examine the relation between ownership structure and the probability of manager replacement in the mutual fund industry. The turnover literature is almost exclusively limited to replacements that occur in public firms primarily because market-based performance metrics are usually not available for privately held firms. This is a shortcoming given the importance of private organizations in most economies.

Using a large sample of mutual funds from 1999 to 2007, we find a positive correlation between public ownership and the likelihood of manager replacement. More importantly, our analysis indicates that publicly traded sponsors are more sensitive to poor performance when deciding to retain or terminate fund managers, and that performance improvements following forced manager replacement are greater in private than public mutual funds. We also find a greater likelihood of fund manager replacement in larger funds within a fund complex suggesting that sponsors monitor managerial performance when funds generate larger revenues. Further testing examining turnovers in team managed funds, where monitoring is a challenge, indicates a greater likelihood of turnovers for public sponsors in team managed funds.

Our findings complement a growing field of literature that examines how differences in information asymmetry and agency problems across public and private organizational structures influence managerial incentives and corporate behavior. Although recent research suggests that organizational structure is an important consideration in many corporate decisions (Asker et al., 2011, Mortal and Reisel, 2011, Michaely and Roberts, 2012, Burgstahler, Hail, and Luez, 2006), more research is needed to increase our understanding as to how organizational form influences mutual funds. We take some initial steps in this direction by demonstrating that sponsor level organizational form matters in the monitoring of mutual fund managers. Several remaining questions center on the labor market for fund managers, such as whether career concerns lead fund managers to favor employment in publicly traded or privately held sponsors and if fund manager compensation schemes vary across organizational forms.

Footnotes
  • 1

    The terms organizational form and organizational structure refer to publicly traded versus privately held ownership of the fund sponsors and not the mutual funds. Similarly, the terms public funds and private funds refer to the ownership type of the fund sponsor. Examples of privately held sponsors include Fidelity, Lord Abbott, Vanguard, and American Funds Group. Putnam, PIMCO (a subsidiary of Allianz AG), and Janus are publicly traded sponsors. The terms fund sponsor and investment company are interchangeable.

  • 2

    Although infrequent, boards can punish sponsors by not renewing the management contract (Khorana, Tufano, and Wedge, 2007).

  • 3

    However, Hwang and Kim (2009) note that the conventional definition of director independence, no financial or family relationships with the chief executive officer (CEO) or the firm, fails to recognize social ties. They also report greater sensitivity of turnover to performance when boards are conventionally and socially independent. Similarly, for mutual funds, sponsors and independent directors are often connected through business interactions (Khunen, 2009). Social and business ties that impede director monitoring may result in weaker turnover-performance sensitivities for private sponsors who have less incentive to discipline underperforming managers.

  • 4

    Although we find 11 sponsors that experience a change in organizational control due to mergers and acquisitons (M&A) activity, only one (Strong Funds) experienced a change in public versus private ownership type. The results are robust to the exclusion of these sponsors.

  • 5

    We compare the data gathered from the CRSP Mutual Fund database with the Morningstar Principia mutual fund database and find 2,382 discrepancies (22% of overall sample). We address this issue by repeating our tests for each database separately and by eliminating any discrepancies. Our results are robust to each treatment. We report results from Morningstar since their data is commonly used by investors and financial media and more consistent than CRSP (Massa et al., 2010).

  • 6

    Although not reported, manager tenure in the control sample of funds not experiencing turnover varies only slightly across organizational type suggesting policies such as mandatory periodic fund manager rotations are similar for public and private sponsors.

  • 7

    Clustered standard errors can be problematic when the number of observations within clusters is small (Donald and Lang, 2007). We repeat the analysis using robust standard errors and find similar results. For completeness, we also compute fund family and time clustered standard errors and find similar results. We avoid coefficient interpretation problems associated with interaction terms in nonlinear models as detailed in Ai and Norton (2003) by including at least one dummy variable in each interaction term.

  • 8

    Each regression model includes performance measures computed over the 24 months prior to the turnover month (t–25, t–1) and over the two years preceding the turnover calendar year end (t–1, t) as well as the 24-month period beginning the year prior to the turnover (t–2, t–1). Our findings are robust to each performance measurement period. Since many of our variables are only available on a calendar year end basis, we report the results for the prior calendar year end performance measures.

  • 9

    Anecdotal evidence from the popular press supports this claim. See, for example, the article entitled “Manager Stability Pays Real Dividends” by Joe Morris in Ignites (June 4, 2007).

  • 10

    We account for potential survivorship bias by noting that the mean public fund replacement year objective adjusted return is about 2% lower than the mean private fund replacement year return (approximately –4% and –2% for public and privates funds, respectively) and that most of the missing observations (about 85%) are for public funds. Since the poor performance and higher incidence of missing public funds biases our private ownership results downward, we also perform multiple imputations to estimate the missing values and find similar results. Further, we employ logit analysis with the dependent indicator variable taking on a value of one if the change in performance is positive and zero if the change in performance is negative or the fund does not survive. The results from the logit analysis concur with the ordinary least square (OLS) results.

  • 11

    In alternate specifications, we account for mean reversion effects by including prior performance measures in place of the forced turnover dummy and find similar results. In additional regressions, we control for potential forward looking bias by including fixed effects for investment objective and year interactions.

  • 12

    We also consider the possibility that serving on multiple boards outside of the fund family impacts directors’ monitoring effectiveness and find similar results.

References

  1. Top of page
  2. Abstract
  3. I. Background and Hypotheses
  4. II. Data and Variables Measures
  5. III. Sample Statistics
  6. IV. Multivariate Analysis
  7. References
  • Adams, J., S. Mansi, and T. Nishikawa, 2010, “Internal Governance Mechanisms and Operational Performance: Evidence from Index Mutual Funds,” Review of Financial Studies 23, 12611286.
  • Ai, C. and E. Norton, 2003, “Interaction Terms in Logit and Probit Models,” Economic Letters 80, 123129.
  • Almazan A., K. Brown, M. Carlson, and D. Chapman, 2004, “Why Constrain Your Mutual Fund Manager Journal of Financial Economics 73, 289321.
  • Anderson, R., S. Mansi, and D. Reeb, 2003, “Founding Family Ownership and the Agency Cost of Debt,” Journal of Financial Economics 68, 263285.
  • Anderson, R., S. Mansi, and D. Reeb, 2004, “Board Characteristics, Accounting Report Integrity, and the Cost of Debt,” Journal of Accounting and Economics 37, 315342.
  • Asker, J., J. Farre-Mensa, and A. Ljungqvist, 2011, “Does the Stock Market Distort Investment Incentives?” NYU and ECGI Finance Working Paper No. 282/2010.
  • Berle, A. and G. Means, 1932, The Modern Corporation and Private Property , New York , NY , Macmillan.
  • Bonnier, K. and R. Bruner, 1989, “An Analysis of Stock Price Reaction to Management Change in Distressed Firms,” Journal of Accounting and Economics 11, 95106.
  • Brav, A., J. Graham, C. Harvey, and R. Michaely, 2005, “Payout Policy in the 21st Century,” Journal of Financial Economics 77, 483527.
  • Brown, K., W. Harlow, and L. Starks, 1996, “Of Tournaments and Temptations: An Analysis of Managerial Incentives in the Mutual Fund Industry,” Journal of Finance 51, 85110.
  • Burgstahler, H., L. Hail, and C. Leuz, 2006, “The Importance of Reporting Incentives: Earnings Management in European Private and Public Firms,” Accounting Review 81, 9831016.
  • Bushee, B., 2001, “Do Institutional Investors Prefer Near-Term Earnings Over Longer-Run Value Contemporary Accounting Research 18, 207246.
  • Campbell, T., M. Gallmeyer, S. Johnson, J. Rutherford, and B. Stanley, 2011, “CEO Optimism and Forced Turnover,” Journal of Financial Economics 101, 695712.
  • Carhart, M., 1997, “On Persistence in Mutual Fund Performance,” Journal of Finance 52, 5782.
  • Chevalier, J. and G. Ellison, 1999, “Career Concerns of Mutual Fund Managers,” Quarterly Journal of Economics 114, 389432.
  • Coffee, J., 1999, “The Future as History: The Prospects for Global Convergence in Corporate Governance and its Implications,” Northwestern University Law Review 93, 641708.
  • Denis, D. and D. Denis, 1995, “Performance Changes Following Top Management Dismissals,” Journal of Finance 50, 10291057.
  • Deuskar, P., J. Pollet, Z. Wang, and L. Zheng, 2011, “The Good or the Bad? Which Mutual Fund Managers Join Hedge Funds Review of Financial Studies 24, 30083024.
  • Ding, B. and R. Wermers, 2009, “Mutual Fund Performance and Governance Structure: The Role of Portfolio Managers and Boards of Directors,” SUNY Albany Working Paper.
  • Donald, S. and K. Lang. 2007, “Inference with Difference-in-Differences and Other Panel Data,” Review of Economics and Statistics 89, 221233.
  • Evans, R., 2010, “Mutual Fund Incubation,” Journal of Finance 65, 15811611.
  • Ferris, S. and X. Yan, 2009, “Agency Costs, Governance, and Organizational Forms: Evidence from the Mutual Fund Industry,” Journal of Banking and Finance 33, 619626.
  • Froot, K., A. Perold, and J. Stein, 1992, “Shareholder Trading Practices and Corporate Investment Horizons,” Journal of Applied Corporate Finance 5, 4258.
  • Furtado E. and M. Rozeff, 1987, “The Wealth Effects of Company Initiated Management Changes,” Journal of Financial Economics 18, 147160.
  • Gaspar, J., M. Massa, and P. Matos, 2005, “Shareholder Investment Horizons and the Market for Corporate Control,” Journal of Financial Economics 76, 135165.
  • Giannetti, M., 2003, “Do Better Institutions Mitigate Agency Problems? Evidence from Corporate Finance Choices,” Journal of Financial and Quantitative Analysis 38, 185212.
  • Hwang, B. and S. Kim, 2009, “It Pays to Have Friends,” Journal of Financial Economics 93, 138158.
  • Jensen, M., 1967, “The Performance of Mutual Funds in the Period 1945-1964,” Journal of Finance 23, 389416.
  • Khorana, A., 1996, “Top Management Turnover: An Empirical Investigation of Mutual Fund Managers,” Journal of Financial Economics 40, 403427.
  • Khorana, A., 2001, “Performance Changes Following Top Management Turnover: Evidence from Open-End Mutual Funds,” Journal of Financial and Quantitative Analysis 36, 371393.
  • Khorana, A., P. Tufano, and L. Wedge, 2007, “Board Structure, Mergers, and Shareholder Wealth: A Study of the Mutual Fund Industry,” Journal of Financial Economics 85, 571598.
  • Khunen, C., 2009, “Business Networks, Corporate Governance and Contracting in the Mutual Fund Industry,” Journal of Finance 64, 21852220.
  • Loderer, C. and U. Waelchli, 2010, “Protecting Minority Shareholders: Listed versus Unlisted Firms,” Financial Management 39, 3357.
  • Massa, M., J. Reuter, and E. Zitzewitz, 2010, “When Should Firms Share Credit with Employees? Evidence from Anonymously Managed Mutual Funds,” Journal of Financial Economics 83, 751792.
  • Michaely, R. and M. Roberts, 2012, “Corporate Dividend Policies: Lessons from Private Firms,” Review of Financial Studies 25, 711746.
  • Michaely, R., R. Thaler, and K. Womack, 1995, “Shareholder Heterogeneity, Adverse Selection, and Payout Policy,” Journal of Finance 50, 573608.
  • Morck, R., A. Shleifer, and R. Vishny, 1989, “Alternative Mechanisms for Corporate Control,” American Economic Review 79, 842852.
  • Mortal, S. and N. Reisel, 2011, “Capital Allocation by Private and Public Firms,” Southern Methodist University Working Paper.
  • Parrino, R., 1997, “CEO Turnover and Outside Succession: A Cross Sectional Analysis,” Journal of Financial Economics 46, 165197.
  • Porter, M., 1992, “Capital Choices: Changing the Way America Invests in Industry,” Council on Competitiveness/Harvard Business School.
  • Qiu, J., 2003, “Termination Risk, Multiple Managers and Mutual Fund Tournaments,” European Finance Review 7, 161190.
  • Rappaport, A., 1990, “The Staying Power of the Public Corporation,” Harvard Business Review 68, 96104.
  • Shleifer, A. and R. Vishny, 1989, “Management Entrenchment: The Case of Manager-Specific Investment,” Journal of Financial Economics 25, 123139.
  • Sirri, E. and P. Tufano, 1998, “Costly Search and Mutual Fund Flows,” Journal of Finance, 53, 15891622.
  • Volpin, P., 2002, “Governance with Poor Investor Protection: Evidence from Top Executive Turnover in Italy,” Journal of Financial Economics 64, 6190.
  • Weisbach, M., 1988, “Outside Directors and CEO Turnover,” Journal of Financial Economics 20, 431460.
  • Yan, X. and Z. Zhang, 2009, “Institutional Investors and Equity Returns: Are Short-Term Institutions Better Informed Review of Financial Studies 22, 893924.