European small-cap fund performance : outperformance of high active share funds with patient strategies

(1)

University of Amsterdam BSc Economics & Business

Bachelor Specialization Economics & Finance

European Small-Cap Fund Performance

Outperformance of High Active Share Funds

with Patient Strategies

Author: Dion Verboom Student Number: 10458751 Thesis Supervisor: Dr. Jan Lemmen Date: January 2018

(2)

‘Competition between asset managers

on the basis of relative performance is inherently

a zero sum game. The asset management

industry can benefit its customers

only to the extent that its activities

improve the performance of companies.’

(3)

Preface and Acknowledgements

At first my thesis topic was active management versus passive management. The aim was to evaluate whether active managed funds outperformed their passive benchmark index. Like many other papers, the outcome probably would have been that active managed funds were not able to outperform their passive benchmark index and active fund manager were not able to create value for their investors. After a discussion with my neighbor, Jeroen Potma, I changed my thesis topic. My neighbor pointed out that it was more interesting to evaluate the reason why active managed funds underperform or outperform their benchmark and referred me to an article of Cremers and Petajisto. This laid the foundation of the thesis that lies before you.

The completion of my thesis would not have been able without several persons. I would like to thank my thesis supervisor, Dr. Jan Lemmen, for guidance throughout my thesis period. I would like to thank Luuk Jagtenberg (Kempen Capital Management) for the assistance collecting data for my research. Special thanks I owe to my neighbor, Jeroen Potma (Kempen), for thinking and reading along, guidance throughout my thesis and discussions regarding active management.

Statement of Originality

This document is written by Dion Verboom who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no

sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work,

(4)

Abstract

This paper evaluates whether European small-cap mutual fund performance improves

as Active Share and Fund Duration increase. Our results show that only high Active

Share funds with patient strategies outperform their passive benchmark. The average

outperformance of high Active Share funds with patient strategies is 1.39 basis points

per year net of costs. This suggests that only managers with enough stock-picking

skills to construct a substantially different portfolio than their passive benchmark and

strong conviction to hold these stocks for a longer period, outperform their

benchmark and are able to add value for their investors. In addition, “the world”

would benefit as well.

Keywords:

Active Management, European Small-Cap Funds, Fund Duration, Mutual Fund Performance, Long-Term

JEL Classifications: G11, G12, G14, G24

(5)

List of Tables

Table 1: Active Share and Fund Duration ranges by Class...18

Table 2: European Small-Cap Sample: Median by Year...20

Table 3: European Small-Cap Sample: Descriptive Statistics...21

Table 4: Active Share and Fund Duration: Four-Factor Alphas of Net Returns...24

Table 5: Active Share and Fund Duration: CAPM Alphas of Net Returns...27

Table 6: Four-Factor Alphas with T-Statistics...32

(7)

List of Figures

Figure 1: Active Management Categories...6

Figure 2: Number of Funds and Stocks in the Small- and Large-Cap Markets...9

Figure 3: Average Number of Analysts per Company...10

Figure 4: Average Research and Development Expenditures...12

Figure 5: Average Job Creation...13

Figure 6: Time period Executives feel most pressure...14

Figure 7: Percentage of Fund by Class... 22

(8)

1. Introduction

Do investors benefit from investing in active managed mutual funds? The majority of previous papers (Jensen, 1967; Malkiel, 1995; Fama and French, 2010) conclude that the Efficient-Market Hypothesis (EMH) holds and active managed mutual funds, on average, underperform their passive benchmark. However, there are some papers (Cremers and Petajisto, 2009; Petajisto, 2013; Cremers and Pareek, 2016) that argue subsets of active managed mutual funds consistently show outperformance. These papers argue that only active managed funds with a substantial different portfolio than their benchmark and a patient investment strategy are able to outperform their benchmark.

First of all, Cremers and Petajisto (2009) find that high Active Share funds outperform, on average, their passive benchmark. They use Active Share to evaluate the performance of active managed United States (US) mutual funds. Active Share describes the share of portfolio holdings that deviates from their passive benchmark. They find that high Active Share funds, on average, outperform their passive benchmark and funds with low Active Share significantly underperform their passive benchmark. Furthermore, Cremers and Pareek (2016) combined Active Share and Fund Duration to evaluate mutual fund performance based on active management and Fund Duration. They find that funds with patient investment strategies outperform frequently trading funds. However, Cremers and Pareek (2016) find that among patient investment funds, only high Active Share funds outperform their passive benchmark.

The aim of this paper is to evaluate whether active European small-cap mutual fund performance improves as Active Share and Fund Duration increase. The European small-cap market may be too information efficient in the short-term for active fund managers to benefit from frequently trading. At the same time, active fund managers with sufficient stock-picking ability to spot mispricing’s in the long-term and strong conviction to hold these stocks for a longer period may be able to capitalize on information inefficiencies in the long-term.

Furthermore, the investment horizons set by active European small-cap funds may influence the performance of invested companies. Shorter investment horizons increase pressure on companies to maximize short-term financial results and increase short-termism at companies. Short-termism is the excessive focus of companies on short-term objectives at the expense of long-term value creation. European small-cap funds with longer investment horizons may decrease the short-term pressure on companies and eliminate short-termism. This could result in better company performance, fund performance and maybe lead to benefits for the society at large (Barton et al., 2017).

The existent literature has not provided research on the relation between active management, Fund Duration and fund performance of European small-cap mutual funds. The main reason is that most mutual fund research is conducted on US mutual funds samples, this is because the available data series of US mutual funds are longer compared to the available data series of European funds. Cremers and Pareek (2016) conducted research on the relation between active management, Fund Duration and fund performance of US mutual funds. One may expect our European small-cap fund sample to perform

(9)

better than Cremers and Pareek’s (2016) US mutual fund sample. This is due to, in general, better performance of European mutual funds (Otten and Bams, 2002; Otten and Schweitzer, 2002) and consistently better performance of small-cap funds compared to large-cap funds (Cremers, 2017).

The research set-up in this paper is as follows. We consider a sample consisting only of active managed all-equity European small-cap funds for the period of 2005-2016. Next, we sort the funds into three classes of Active Share and Fund Duration, resulting in nine portfolios. Subsequently, we use the Carhart (1997) Four-Factor regression model to evaluate the equally-weighted quarterly net returns of the nine portfolios separately. Thereafter, we compare the performance of the portfolios along both sides of Active Share and Fund Duration. For robustness, we also consider the CAPM regression.

Among all portfolios, only high Active Share funds with patient investment strategies outperform, on average, their benchmark. These funds are in the top classes of both Active Share and Fund Duration and are defined as substantially different in stock holdings than their

benchmark and have long investment horizons. These funds outperform their passive benchmark by 1.39 basis points per year (T-Statistic of 1.62). On the other hand, funds with low Active Share and with short investment horizons underperform their passive benchmark by -0.45 basis points (T-Statistic of 0.62). This suggests that only managers with sufficient stock-picking skills to

construct a substantial different portfolio than their benchmark and have strong conviction to hold these stocks for a longer period, are able to outperform their benchmark and create value for their investors.

Furthermore, we find that high Active Share funds in itself are not associated with

outperformance. Our results show that high Active Share funds, on average, underperform their passive benchmark by -0.60 basis points (T-statistic of 1.21). Only high Active Share funds with patient strategies outperform their benchmark. Rather, we find no evidence that high Active Share funds perform better than low Active Share funds. The high Active Share fund portfolio outperformed the low Active Share fund portfolio by an economically small and statistically insignificant abnormal return of 0.14 basis points (T-Statistic of 0.23).

Moreover, we find that long Fund Duration itself is not associated with outperformance. The long Fund Duration portfolio underperformed their passive benchmark by -0.21 basis points per year (T-Statistic of 0.26). Only long Fund Duration funds with substantially different portfolios than their benchmark outperformed their passive benchmark. However, we do find evidence that long Fund Duration funds outperform short Fund Duration funds. The long Fund Duration portfolio outperformed the short Fund Duration portfolio by an economically modest 0.95 basis points per year (T-Statistic of 1.09). This shows that funds with patient investment strategies perform better than frequently trading funds. Funds with short Fund Duration underperform their passive benchmark, despite of Active Share class.

To conclude, our results show that as Fund Duration increases, European small-cap mutual fund performance improves. However, among long Fund Duration funds, only high Active Share funds

(10)

outperform their benchmark. We find no evidence for high Active Share outperformance in general. Only among long Fund Duration funds, high Active Share improves European small-cap mutual fund

performance. This suggests that only managers with stock-picking skills to construct a substantial different portfolio than their benchmark and a strong conviction to hold these stocks for a longer period are able to add value for their investors. At the same time, the underperformance of short Fund Duration funds suggest that the European small-cap market is too information efficient for asset managers to benefit from frequently trading. Finally, our results suggest that funds with longer investment horizons reduce pressure on invested companies and improve company performance, fund performance and the economy at large. This suggest that when funds commit to a long-term approach companies, funds and “the world” would benefit.

(11)

2. Literature review

2.1 Contradictory results of mutual fund performance

The performance of active fund managers is a much debated subject in research papers. The results of these papers are contradictory. The majority of research papers (Jensen, 1967; Malkiel, 1995; Fama and French, 2010) conclude that active mutual funds, on average, underperform their passive benchmark, either due to high costs or a lack of stock-picking skills. However, some research papers (Otten and Bams, 2002; Cremers and Petajisto, 2009; Cremers and Pareek, 2016) argue that subsets of mutual funds continuously show outperformance. Furthermore, several papers (Otten & Bams, 2002; Otten and Schweitzer, 2002) found that European mutual funds perform better than their US counterparts.

First of all, Wermers (2000) evaluates the value of active fund management. He examines the performance of US equity funds and decomposes the fund performance into stock-picking ability, characteristic timing, style, transaction cost and expenses. Wermers (2000) found that on a gross return basis, US equity funds outperformed their passive benchmark by 1.30 basis points. However, after deducting costs, mutual funds underperformed their passive benchmark by 1.00 basis point. Wermers (2000) found that part of the difference between gross and net return performance is due the lower average return of nonstock holdings and the remaining part due transaction costs and fees. Wermers (2000) concludes that considering only stock holdings, active fund managers add enough value to cover their transaction costs and expenses.

In contrast to Wermers (2000), Fama and French (2010) found in their paper that only a few active managed mutual funds have enough skill to cover their costs. They evaluate US mutual fund performance from 1984 to 2006 and examine whether active fund managers outperform their benchmark. To measure the fund performance, they use the Fama & French three-factor model (1993) and the Carhart four-factor model (1997) to control for market, size, style and momentum characteristics. Fama and French (2010) found that, on aggregate, mutual funds underperform their passive benchmark by the amount of expenses they charge investors. They argue that the mutual funds underperformance is due to higher fees active managed funds charge to investors and due to higher transaction costs compared to their passive benchmark.

Instead of evaluating the US mutual fund industry, Otten and Bams (2002) examine the

performance of 506 European mutual funds from the five most important European countries. They use the Carhart four-factor model (1997) to measure the risk-adjusted fund performance. They found that four out of the five countries exhibit a positive alpha on a net return basis with only the United Kingdom (UK) showing a significant outperformance. They also examine the gross return of the European mutual funds. They found that four out of the five countries exhibit a significant outperformance. Only German mutual funds underperformed their passive benchmark. The results of Otten and Bams (2002) deviate from most US studies (Wermers, 2000; Fama and French, 2010) that argue mutual funds underperform their

(12)

could be the smaller market importance of the European mutual fund industry compared to their US counterparts. European mutual funds hold a substantial less amount of domestic equity than US mutual funds. This makes it easier for European mutual funds to outperform as a group. In their research Otten and Bams (2002) also found that especially small-cap funds were able to add value.

Further, Otten and Schweitzer (2002) compare the characteristics and performance of the European and US mutual funds industry. They found that the European mutual fund industry is still lagging the US mutual fund industry in total asset size and market importance. However, the number of funds in Europe is larger than the US mutual fund industry, which means European funds are smaller on average. They found that the expense ratios in the US are, on average, slightly higher than European mutual funds, respectively 1.4% for US funds and 1.2% for European funds. Finally, Otten and Bams (2002) compared the performance of the two mutual fund industries. They found that European mutual funds perform better than US mutual funds, and that small-cap funds outperformed their passive benchmark and all other funds in both industries. They argue that the lower fees in Europe are an important factor for the better performance of European mutual funds.

2.2 Survivorship bias

When a sample only includes current existing funds and exclude merged and liquidated funds,

survivorship bias arises. Malkiel (1995) examines the performance of mutual funds and the magnitude of survivorship bias between 1971 and 1991. He found that samples only including current existing mutual funds, significantly overstate the performance of mutual funds and understate investors risk. Since bad performing funds usually are liquidated or merged, only better performing funds survive and are included in the samples. This creates an upward bias of the mutual fund performance. Malkiel (1995) found in his research that samples only consisting of current existing funds have an insignificant negative alpha. This implies that even surviving funds can’t add enough value to cover their expenses. Further, he found that samples including all mutual funds that exist throughout the sample period, on aggregate, tend to underperform their passive benchmark net and gross of expenses. Malkiel (1995) concludes that the results show that the security market is efficient. He adds that investors are better off buying low cost index funds, then try to select an active manager that persistently outperforms the market.

Moreover, Elton, Gruber and Blake (1996) also examine the performance of mutual funds and the impact of survivorship bias on mutual fund performance. They state that excluding mutual funds creates survivorship bias, because funds that are excluded are likely bad performing funds. The remaining sample of mutual funds overstates the mutual fund performance and understates investors risk. Further, they found that the majority of funds that perform poorly are merged into other funds. Hereby, burying the funds poor performance records into the merged funds. Elton, Gruber and Blake (1996) use two samples to examine the impact of survivorship bias on mutual fund performance. One survivorship-bias free sample and one sample that excludes liquidated and merged funds. They found just as Malkiel (1995) that the sample with survivorship bias overstates the performance of mutual funds. They conclude that the

(13)

survivorship bias becomes larger as the length of the sample period increases. The results of these papers (Malkiel, 1995; Elton, Gruber and Blake, 1996) state that survivorship bias is an important factor to consider when examining mutual fund performance.

2.3 Outperformance subsets of mutual funds 2.3.1 Active management

As mentioned, most papers (Malkiel, 1995; Wermers, 2000; Fama and French, 2010) conclude that mutual funds, on average, underperform their passive benchmark. However, there are several papers (Cremers and Petajisto, 2009; Petajisto, 2013; Cremers and Pareek, 2016) that argue subsets of mutual funds consistently show outperformance. These papers focus more specifically on the relationship between active management and performance of mutual funds. These papers argue that a necessary condition for outperformance is a substantially different portfolio then the passive benchmark.

Cremers and Petajisto (2009) study the relationship between types of active management and mutual fund performance in the US. They introduce Active Share as a new measure to quantify active management. Active Share describes the share of portfolio holdings that deviates from the benchmark index. They combine Active Share with tracking error, which represents the difference between the return volatility of a fund and the benchmark, to sort mutual funds into various active management categories. The various categories are stock pickers, closet indexing, concentrated funds and factor bets.

Figure 1. Active Management Categories

(14)

Compared to previous papers, Cremers and Petajisto (2009) differ in their approach by focusing more specifically on the types of active management. Most previous papers (Wermers, 2000; Fama and French, 2010) focus on mutual fund performance directly. They examine performance of all mutual funds as one group. Cremers and Petajisto (2009) refine those results by sorting mutual funds into different active management categories to get more specific results.

Finally, Cremers and Petajisto (2009) examine the performance of the various active management categories separately. They found that funds in the highest Active Share categories outperformed their benchmark before and after expenses. The active management categories with the lowest Active Share underperformed their passive benchmark. Further, they found no relation between tracking error and mutual fund performance. This means that stock pickers and concentrated funds are the only categories that show outperformance. Both categories hold funds with high Active Share values. Following the research method of Cremers and Petajisto (2009), Petajisto (2013) added six more years to his research from 2004 till 2009. He found the same results as Cremers and Petajisto (2009) except for the

concentrated funds. This group suffered hard during the 2004-2009 period. Both papers (Cremers and Petajisto, 2009; Petajisto, 2013) conclude that there are inefficiencies in the security market that active managers with stock-picking skills can exploit.

2.3.2 Patient strategies

Cremers and Pareek (2016) study the relation between holding duration and mutual fund performance of US active managed funds. They use Active Share (Cremers and Petajisto, 2009) and three holding duration proxies to sort funds into different portfolios. The first proxy is Fund Duration, which measures the length of time a fund holds a security in their portfolio over the last five years. The other two proxies they use are holding-bases fund turnover and the self-declared fund turnover ratio. They use the

benchmark-adjusted returns to compare mutual fund performance along both dimensions of Active Share and holding duration.

The result of Cremers and Pareek (2016) show that among all portfolios, only high Active Share funds with long Fund Duration outperform their passive benchmark. High Active Share and long Fund Duration portfolios are defined as substantially different from their passive benchmark in portfolio holdings and have patient investment strategies. They found that funds trading

frequently and have short Fund Durations systematically underperform their benchmark, regardless of Active Share value. Further, they found that patient managed funds in itself are not related to outperformance. Only patiently managed funds with high Active Share show

outperformance.

Based on these results, Cremer and Pareek (2016) argue that the stock market is too information-efficient and too competitive for funds to systematically benefit from frequently short-term trading. At the same time, fund managers who are able to spot mispricing in the long-term and have strong convictions can add value in the long-term. The reason for managers to have a strong conviction is the possibility that

(15)

trading on long-term mispricing’s in the short-term becomes worse. The fund managers risk to

underperform in the short-term before trading on long-term mispricing pays off (Cremers, 2016). Only active fund managers with enough stock-picking skills to construct a portfolio that is substantially different then their passive benchmark and have a strong conviction to hold these stocks for a longer period are able to outperform their passive benchmark and add value.

(16)

3. Trends in the Mutual Fund Market

3.1 Small-cap premium

Small-cap funds, on average, have better performance records than mid- and large-cap funds. This is because small capitalization stocks, in general, perform better than larger capitalization stocks, this is referred to as the small-cap premium. Active fund managers need sufficient investment opportunities to outperform their passive benchmark and create value for their investors. Investment opportunities emerge from market inefficiencies. Therefore, less efficient markets will offer more valuable investment

opportunities and have more potential for active managers to create value. As a result, small-cap managers have greater investment opportunities than larger cap managers, due the greater inefficiencies in the small-cap market.

Cremers (2017) evaluates the performance difference between small- and large-cap funds. He found that small-cap funds, on average, outperformed large-cap funds. Cremers (2017) argues that these results suggest that small-cap fund managers have greater investment opportunities than large-cap fund managers. There are significantly less funds and substantially more stocks in the small-cap market compared to the large-cap market. This gives small-cap managers a broader and fragmented range of investment opportunities.

Figure 2. Number of Funds and Stocks in the Small- and Large-Cap Markets

Figure 2 reports the number of funds and stocks in both the small- and large-cap market. The bars on the left side represent the number of small- and large-cap funds. The right bars represent the number of stocks in the small- and large-cap market. Green bars represent small-caps and grey bars represent large-caps.

Source: Kempen Capital Management.

0 500 1000 1500 2000 2500 3000 3500 4000 4500 Number of funds Number of stocks Global small caps Global large caps

(17)

Furthermore, small-cap markets have more investment opportunities due to greater information

inefficiencies. The reason for the information inefficiencies is that small-cap stocks are under researched. This is because investment banks, who conduct the majority of investment research, conduct less research on small-cap stocks. The average analysts per company is substantially lower for small-cap stocks than large-cap stocks. Investment banks conduct less research on small-cap stocks because the earnings on small-cap research are less than research on large-cap stocks. Small-cap stocks are, in general, less liquid and have less need for financing than larger cap stocks. As a result, small-cap stocks are less traded and investment banks earn less on small-cap research. Moreover, because investment banks earn less on small-cap research, small-cap stocks are often used as “training material” for young and inexperienced analysts. Consequently, research on small-cap stocks is often of moderate quality. The moderate quality and limited amount of research creates information inefficiencies in the small-cap market. These inefficiencies create investment opportunities for small-cap managers with stock-picking abilities to exploit.

Figure 3. Average number of Analysts per Company

Figure 3 reports the average number of analyst per company ranked on market capitalization. The green bars represent small-cap stocks and the grey bars represent large-cap stocks.

Source: Kempen Capital Management.

Moreover, the impact of the new European Union guideline, Markets In Financial Instruments Directive II (MIFID II), could increase the information inefficiencies in the European small-cap market even further. MIFID II is a new implemented guideline from the European Union to increase transparency

0 5 10 15 20 €20 m-€1bn _€1b-€4b _€4b-€20b _>€20b Global small caps Global large caps

(18)

of markets and strengthen investors protection (Schumacher, 2018). In the new rule setup of MIFID II, asset managers are not allowed to pay investment banks through trade commission. Instead, asset managers need to make explicit payments for investment research. Almost all European asset managers absorb these research costs and pay for research themselves (Mooney, 2018). As a result, asset managers will be more selective in research from investment banks. Consequently, asset managers reduce the number of investment banks they buy research from and this reduces the amount of research available. This results in more inefficiencies, which managers with stock-picking skills can exploit.

Further, active small-cap managers also gain a valuable advantage from the small-cap benchmark characteristics (Pattinson, 2009). In order to outperform their passive benchmark, mutual funds need to be sufficiently different in stock holdings than their passive benchmark. As Cremers and Petajisto (2009) and Cremers and Pareek (2016) documented in their papers, high Active Share is correlated with

outperformance. Small-cap benchmarks are more diversified and evenly distributed then large-cap benchmarks. Indeed, large-cap benchmarks tend to be top heavy concentrated (Pattinson, 2009). This means that the largest stocks in the large-cap benchmarks consume most of the benchmark portfolio. This makes it inherently easier for cap managers to construct a high Active Share portfolio. The small-cap benchmark characteristics gives small-small-cap managers another advantage to create value for their investors.

Finally, passive management can also add to market inefficiencies, particular in small-cap markets. Passive index funds buy and sell stocks as quickly as possible to follow the market. When investors buy these index funds, there is no discrimination among stocks (Pattinson, 2009). Therefore, buying and selling to rebalance these passive portfolios creates inefficiencies at margins. One such an inefficiency is that an increase in passive trading leads to a higher correlation between individual stocks. This creates investment opportunities for active managers with stock-picking ability to create value.

3.2 Short-termism

In recent years publicly listed companies had an excessive focus on short-term performance at the expense of long-term value creation. This is referred to as short-termism. Companies that experience short-termism delay or cancel new valuable investments to hit short-term targets and thereby sacrifice value in the long-term. The main source of short-termism is the pressure on companies from financial markets. Asset owners, who invest in mutual funds, and asset managers, who manage mutual funds, set short investment horizons and this pressures companies to maximize short-term performance. In 2013, BlackRock, Canada Pension Plan Investments Board (CPPIB) and McKinsey & Company created the Focusing Capital on the Long Term (FCLT Global) initiative with the aim of conducting research on long-termism and developing tools and solutions to secure the right balance between short- and long-term performance. A survey conducted by McKinsey Global Institute (MGI) shows that the pressure company executives feel increased over the past years (Barton et al., 2016). This increases short-termism at companies and makes companies less able to invest and create value in the long-term. Barton and

(19)

Wiseman (2014) of the FCLT Global initiative, argue that a crucial breakthrough to reverse short-termism would be if asset owners and asset managers shift from quarterly capitalization to long-term value

creation and engage in active ownership.

The problem of short-termism is that companies sacrifice value in the long-term to hit short-term targets, thereby reducing company performance in the long run. Research of FCLT Global shows that short-term orientated companies underperform their long-term counterparts on a range of financial and economic metrics (Barton et al., 2017). They use systematic company level measurements of long- and short-termism to separate short- and long-term companies. First of all, Barton et al.’s (2017) findings show that companies with a long-term approach, outperform short-term orientated companies on some key fundamentals of value creation. They found that long-term orientated company revenues, earnings and economic profits increased more than short-term companies, because long-term companies maintained consistent and sustainable investment strategies. Moreover, long-term companies invested 46% more in research and development (R&D) than short-term companies to create value in the long-term. This is consistent with the longer investment horizon of long-term orientated companies. As a result, long-term companies experienced higher market capitalization and a greater total return to shareholders than short-term companies.

Figure 4. Average Research and Developing Expenditures

Figure 4 reports the average R&D expenditure of companies. The grey line represents long-term orientated companies and the blue line the short-term orientated companies. The blue zone represents the Global Financial Crisis period.

(20)

Besides the financial benefits, long-term companies also improve the economy at large. In their paper, Barton et al. (2017) found that long-term orientated companies created, on average, 12,000 more jobs than other companies over a 15-year period. They argue that when all US companies performed on a long-term approach, this would have added 5 million extra jobs in the US. The value of the 5 million jobs exceeds $1 trillion and could add more than 0.8% to the US GDP per year. Barton et al. (2017) conclude that longer strategic time horizons add value for companies and their investors.

Figure 5. Average Job Creation

Figure 5 reports the average job creation of companies. The blue line represents the long-term orientated companies and the grey line short-term orientated companies.

Source: McKinsey Corporate Performance Analytics; S&P Capital IQ; McKinsey Global institute Analysis.

The main source of short-termism is the pressure from financial markets. Executives and board members feel pressure from investors to deliver short-term financial results. As a result, executives set short-term targets and strategies to maximize short-term performance and communicate short-term metrics to investors. Subsequently, the short-term metrics shorten investors time horizons and this results in a negative feedback loop. In 2016 MGI conducted a survey of more than thousand board members and c-suite executives about their company’s strategic time horizon (Barton et al., 2016). In the survey, the executives reported that their strategic time horizon is too short. Indeed, half of the respondents claim they use a strategic time horizon of less than three years, while they argue they should use a time horizon of more than three years. As result, executives feel they are investment restrained and are not able to create value in the long-term. As Figure 6 shows, executives especially feel pressure to deliver financial results within two years. Even worse, the MGI survey showed that the short-term pressure on executives had increased over the past five years and executives and board members expect short-term pressure

(21)

continuous to increase. Furthermore, the 2016 MGI survey reports that companies struggle to keep long-term goals their top priority. Among the executives from short-long-term orientated companies, 47% report they would delay new valuable investments if this results in missing the next quarterly earnings by $1 cent, even if this sacrifices value in the long-term. Almost all the respondents from the MGI survey conclude that longer time horizons would increase company performance and returns for investors.

Figure 6. Time period Executives feel most pressure

Figure 6 reports the percentage in which time period executives and board members feel the most pressure to deliver financial performance. Respondents without an answer are reported in the “don’t know” class.

Source: CPPIB McKinsey Focusing Capital on the Long-term 2016 survey.

As mentioned before, the main problem to reverse short-termism is the pressure from financial markets to maximize short-term financial results. Barton and Wiseman (2014) argue that a breakthrough would be if asset owners and asset managers shift from quarterly capitalization to long-term value

creation. Thereby, engaging in active ownership and influence companies to pursuit long-term objectives. First of all, Barton and Wiseman (2014) argue that asset owners should clearly define long-term

objectives and their risk appetite. Asset owners should define clearly how much potential downside they tolerate, how much short-term underperformance is tolerated and ensure the portfolio is invested in line with the long-term investment horizon and risk objectives. It is crucial that the assets are managed in line with the long-term objectives set by asset owners. However, asset managers may have different

investment horizons and objectives than asset owners. This creates agency issues. To ensure that asset managers’ behavior is in line with the asset owners’ objectives, customized mandates between asset owners and managers can be used to minimize these agency issues. Next, asset managers need to implement the long-term objectives set by asset owners in the invested companies. Barton and Wiseman

(22)

(2014) argue that asset managers should engage in active ownership and continuous communication to build a long-term relationship with the companies they invest in. Hereby, asset managers can influence companies to pursuit the long-term objectives and reduce the pressure on company executives. Asset managers should focus on companies instead of securities to create long-term shareholder value. Finally, asset managers should demand long-term metrics from companies. Companies should provide investors with long-term metrics instead of quarterly earnings. Hereby, investors are updated with the process to achieve the long-term objectives, instead of the process to achieving short-term objectives. This eliminates the asset owner’s incentive to pressure companies in short-term financial results. Barton and Wiseman (2014) conclude that when asset owners and asset managers commit to a long-term approach and clearly communicate the objectives, asset owners, asset managers, companies and “the world” benefits.

(23)

4. Data, Methodology and Descriptive Statistics

4.1 Data

4.1.1 European Small-Cap Fund Sample

The data used in this paper is obtained from the Factset mutual fund database. The funds included in our sample are actively managed all-equity European small-cap funds from developed countries for the period 2005-2016. The sample includes all mutual funds that existed during the sample period of 2005-2016. This includes merged and liquidated funds. Our final sample consists of 137 actively managed European small-cap funds.

From the Factset database we used the quarterly net fund returns. These are the returns after costs such as fees, transaction costs and other expenses. The returns from United Kingdom funds are first converted from Pounds to Euros. To calculate Active Share and Fund Duration, we used the annual stock holdings.

The benchmark to which we compare the European small-cap mutual funds is the MSCI Europe small-cap index. This index represents the small capitalization companies from the 15 developed countries in Europe. The quarterly net index returns and the annual stock holdings of the MSCI Europe small-cap index are also obtained from the Factset database.

4.1.2 Sample Selection

To collect only all-equity European small-cap funds in our sample, the following sample selection criteria are used. First, we only include actively managed mutual funds that are not focused on a specific sector and exclude index funds, sector funds and ETFs. Second, we only include funds primarily investing in small capitalization companies in developed European markets and exclude global funds. Further, we exclude funds with return series shorter than two years to get a representative net return sample. Next, only funds with at least two years of stock holdings are included. This is because we need at least two years of stock holdings to calculate Fund Duration. Finally, for funds with multiple share classes, only the capitalization classes are included and the distribution classes are excluded. This is because the

distribution classes pay out dividend that is not incorporated in the net fund return. This makes the

distribution class returns not representative. The capitalization classes reinvest the capital gains, so that all capital gains are incorporated in the net fund return. The remaining funds are included in our sample.

When using sample selection criteria to create a sample biases can arise. As mentioned before, Malkiel (1995) and Elton, Gruber and Blake (1996) found in their papers that only including funds with data for the entire sample period creates survivorship bias. They found that survivorship bias

overestimates mutual fund performance and conclude that it is an important factor to consider. To control for survivorship bias in this research, all mutual funds that existed in the sample period are included. This includes merged and liquidated funds. Due to our sample selection criteria our sample size reduced almost by half from 237 to 137 funds. Especially the criteria of at least two years return series and stock

(24)

holdings reduced the sample size. The exclusion of these funds creates sampling bias. Due to the comprehensive data requirements our sample period starts in 2005 and ends in 2016.

As mentioned before, the benchmark we used to compare the mutual funds is the MSCI Europe small-cap index. The MSCI Europe small-cap index is a representation of small-cap equities of the 15 developed European countries. This index is the best candidate to compare the mutual funds because the index represents small-cap equities of developed European countries and our sample only includes small- cap funds in developed European countries. The factors of the four-factor model and the risk free rate which are used to evaluate mutual fund performance are obtained from Ken French’s website.

4.2 Methodology

4.2.1 Active Share and Fund Duration

Cremers and Petajisto (2009) introduced Active Share as a new measurement of active management. Active Share measures the size of active positions of a portfolio and describes the share of portfolio holdings that deviates from the benchmark holdings. Cremers (2017) introduced a new formula for Active Share. This formula is as follows:

𝐴𝑐𝑡𝑖𝑣𝑒 𝑆ℎ𝑎𝑟𝑒 = 100% − 𝑀𝑖𝑛 𝑊4567,9, 𝑊:;6<=>?@A,9 B

9CD

𝑑 𝑊4567,9 > 0

where N is the total number of stocks in a fund’s portfolio. The variable indicator on the right-hand side is 1 for positive stock positions and otherwise 0. When the weight of a stock is positive the minimum of each stock’s weight in the fund, 𝑊4567,9, and in the benchmark, 𝑊:;6<=>?@A,9, is the overlapping weight between the fund and the benchmark. This new formula expresses Active Share as 100% minus the overlapping weights between the fund and the benchmark. Consequently, Active Share is only lowered by overlapping positions between the fund and the benchmark. Funds with no overlap with the benchmark have an Active Share of 100% and funds with identical portfolios as the benchmark have an Active Share of 0%.

Next to Active Share, Fund Duration is introduced by Cremers & Pareek (2016). Fund Duration measures the period stocks are held continuously in a fund’s portfolio and is an indicator of fund’s investment horizon. The formula is as follows:

𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛_9,J,KLD = 𝑑_9,J,KLD= 𝑇 − 𝑡 − 1 𝛼9,J,O 𝐻_9,J,O+ 𝐵_9,J,O + 𝑊 − 1 𝐻_9,J,O 𝐻_9,J,O+ 𝐵_9,J,O KLD OCKLS

where W is the period for which Duration measured how long a stock is continuously held by a fund. Cremers and Pareek (2013) choose 20 quarters for W, because beyond that period the effect would seem marginal. In our research we use the same time period, only we use years instead of quarters. The total percentage of a stock bought by a fund between the current year, T, and 5 years backwards is 𝐵_9,J,O. Further, the total percentage of a stock outstanding by a fund 5 years backwards is 𝐻9,J,O. Finally, the

(25)

percentage of a stock bought or sold by a fund between years during the 5 year time period is 𝛼9,J,O, where 𝛼9,J,O > 0 for buys and 𝛼9,J,O < 0 for sells. When a stock is not included in a fund 1 year backwards from the current year, 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛9,J,KLD = 0. For the calculation of Fund Duration, we need at least 2 years of fund holdings. Funds with less than 2 years of fund holdings are excluded from our sample. We calculate the Fund Duration by averaging 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛_9,J,KLD of all stocks in a fund’s portfolio.

4.2.2 Performance evaluation

The main purpose of this research is to evaluate whether European small-cap mutual fund performance improves as Active Share and Fund Duration increase. To evaluate the European small-cap mutual fund performance based on active management and fund investment horizon, for each fund we calculate both Active Share and Fund Duration at the end of each year in our sample period. Next, we sort each year the funds into three classes of Active Share and Fund Duration. Table 1 shows the ranges for each class of Active Share and Fund Duration. The bottom Active Share class consists of funds with an Active Share lower then 90% and the top class higher then 95%. The funds in the bottom class of Fund Duration hold stocks less than 0.75 year (9 months). The funds in the top class of Fund Duration hold stocks longer than 1.75 years (21 months).

Table 1. Active Share and Fund Duration ranges by Class

Table 1 reports the class ranges for both Active Share and Fund Duration. The ranges are reported for each separate class for both variables. Active Share is given in percentages and Fund Duration in years.

T11 T2 T3

Active Share < 90% 90% - 95% > 95% Fund Duration < 0.75 0.75 - 1.75 > 1.75

Thereafter, we combine Active Share and Fund Duration to create 9 equally-weighted portfolios.

Next, we use the Carhart (1997) Four-Factor regression model to evaluate the European small-cap mutual fund performance. We evaluate the equally-weighted quarterly net returns of the 9 portfolios. The Carhart (1997) Four-Factor is the standard used regression model in academic papers to evaluate fund performance. The regression model controls for any exposure to market, size, value and momentum factors. The formula of the Carhart (1997) Four-Factor regression model is as follows:

1

(26)

𝑅9,O = 𝛼9+ 𝛽D,9(𝑀𝐾𝑇)O+ 𝛽Y,9(𝑆𝑀𝐿)O+𝛽[,9(𝐻𝑀𝐿)O+ 𝛽\,9(𝑀𝑂𝑀)O+ 𝜀9

where, 𝛼9, alpha indicates a possible outperformance or underperformance of a portfolio. The dependent variable on the left-hand side of the formula, 𝑅_9,O, is portfolio return in excess of the risk-free rate. The risk-free rate is the 1-month US treasury bill and is obtained from the Kenneth French website. Further, the first independent variable on the right-hand side, 𝑀𝐾𝑇O, is the Market factor. This is the MSCI Europe small-cap index return in excess of the risk-free rate. Next, the second independent variable, 𝑆𝑀𝐿_O, is the Small-Minus-Big factor and measures the excess return of small capitalization stocks over large

capitalization stocks. The third independent variable, 𝐻𝑀𝐿O, is the High-Minus-Low factor and measures the excess return of value stocks over growth stocks. The last independent variable, 𝑀𝑂𝑀_O, is the

Momentum factor and measures the excess return of past winners over past losers. Furthermore, the error term, 𝜀₉, is the residual of the regression. Finally, the betas, 𝛽_D,9,𝛽_Y,9,𝛽_[,9 𝑎𝑛𝑑 𝛽_\,9, represent the volatility of the portfolio to the four factors. The data of the factors, except for the Market factor, is obtained from the Kenneth French website. For robustness of this research, we also conduct the one factor Capital Asset Pricing Model regression. This regression only consists of the Market factor.

4.3 Descriptive Statistics

Table 2 reports the annual descriptive statistics of our European small-cap mutual funds sample. The median of the funds, stocks, Active Share and Fund Duration are reported by year. At the start of our sample period the number of funds is equal to 81. The number of funds rises to 93 in 2012 and equals 83 at the end of our sample period in 2016. The median number of stocks included in a fund is stable around 75 stocks. Further, the median Active Share of each year in our sample period is higher than 90%, with the highest median of 99% in 2008. This is in line with the perception that small-cap funds typically have a high Active Share value. Finally, the Fund Duration median continuously rises from 0.69 in the begin of our sample period to 1.36 in 2012 and remains fairly stable thereafter. The increase of the Fund Duration median could be due to the data collection of Factset. There is less available data of fund stock holdings at the start of our sample period in the Factset database. Consequently, this decreases the Fund Duration of funds and induces the low Fund Duration medians at the start of our sample period.

(27)

Table 2. European Small-Cap Mutual Fund Sample: Median by Year

Table 2 reports the annual medians of the mutual fund sample. The medians for the variables are reported for each sample year.

Year # of Funds # Stocks Active Share Fund Duration 2005 81 90 94% 0.69 2006 83 85 95% 0.88 2007 76 77 96% 0.92 2008 80 77 99% 0.99 2009 80 89 94% 1.00 2010 80 82 93% 1.08 2011 81 79 93% 1.17 2012 93 75 93% 1.36 2013 87 75 93% 1.32 2014 83 66 93% 1.20 2015 77 73 94% 1.27 2016 83 78 94% 1.25

Panel A of Table 3 reports the basic descriptive statistics of the entire sample. The number of observations, mean, standard deviation, minimum and maximum are given for Active Share, Fund Duration and number of stocks included in a fund. The Active Share mean of the entire sample is 93% and standard deviation 7%. This shows that the funds Active Share values tend to be close to the mean. The lowest Active Share value is 52% and the highest 100%. The Fund Duration values have a mean of 1.18 and a standard deviation of 0.53. This shows that the Fund Duration values are spread out over a wider range of values than Active Share values. Fund Duration has a minimum of 0.01 and a maximum of 3.44. The number of stocks in a fund have a mean of 116, standard deviation of 161, a minimum of 12 and a maximum of 1269.

Panel B of Table 3 reports the correlation between Active Share, Fund Duration and number of stocks in our sample. None of the three variables is highly correlated with another variable, with the highest correlation equal to -13% for Active Share and Fund Duration. The correlation with the number of stocks for Active Share and Fund Duration are respectively 9% and -2%.

Panel C of Table 3 reports the average percentage of funds in each of the nine portfolios and the three classes of Active Share and Fund Duration. The middle classes of Active Share and Fund Duration hold the most funds. The middle class of Active Share hold 40.3% of the funds and the middle class of Fund Duration 58.6%. Further, it shows that of the highest Fund Duration class, almost half of the funds

(28)

also are in the top Active Share class. Which means that many funds with long Fund Durations are also very active.

Table 3. European Small-Cap Mutual Fund Sample: Descriptive Statistics

Panel A reports the descriptive statistics for the European small-cap mutual fund sample used in this paper. Panel B reports the correlation between variables of the European small-cap mutual fund sample. Panel C reports the percentage of funds in each of the 9 portfolios based on Active Share and Fund Duration. The portfolios are obtained by first sorting the funds into groups of Active Share and then sorted by Fund Duration in each of the Active Share groups. The European small-cap mutual fund data is obtained from the Factset mutual fund database. Panel A

# of obs Mean StDev. Min Max

Active Share 671 93% 7% 52% 100% Fund Duration 671 1.18 0.53 0.01 3.44 # Stocks 77,883 116 161 12 1269

Panel B

Active Share Fund Duration # Stocks

Active Share 100% Fund Duration -13% 100% # Stocks 9% -2% 100% Panel C Active Share T1 T2 T3 Sum T1 5.4% 9.8% 7.9% 23.2% T2 13.4% 24.4% 20.8% 58.6% T3 3.3% 6.1% 9.8% 19.2% Sum 22.1% 40.3% 36.9% Fu nd D ur at io n

(29)

As mentioned before, the funds in our sample are sorted into three classes of Active Share and Fund Duration. Panel A of Figure 7 shows the percentage of funds in each class of Active Share by year and Panel B of Figure 7 shows the percentage of funds in each class of Fund Duration by year.

Figure 7. Percentage of Funds by Class

Panel A of Figure 7 shows the percentage of funds in each of the three Active Share classes by year. The blue zone represents the lowest class of Active Share with values below 90%. The pink zone represents the middle class of Active Share with values between 90% and 95% and the green zone represents the highest Active Share class with values higher than 95%. Panel B of Figure 7 shows the percentage of funds in each of the three Fund Duration classes by year. The blue zone represents the lowest Fund Duration class with values below 0.75. The pink zone represents the middle class of Fund Duration with values between 0.75 and 1.75. The green zone represents the highest Fund Duration class with values higher than 1.75.

Panel A. Percentage of funds by Active Share class

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Year

> 95% 90% - 95% < 90% Pe rc en ta ge of fu nd s

(30)

Panel B. Percentage of funds by Fund Duration class 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Year

< 0.75 0.75 - 1.75 > 1.75 Pe rc en ta ge of fu nd s

(31)

5. Active Share, Fund Duration and Fund Performance

5.1 European Small-Cap Mutual Fund Performance

To evaluate whether European small-cap mutual fund performance improves as Active Share and Fund Duration increase, we sorted all funds in our sample into 9 portfolios based on Active Share and Fund Duration. Using quarterly net returns, we evaluate separately the performance of the 9 portfolios. Table 4 reports the annualized Four-Factor alphas of the equally-weighted portfolios.

Table 4. Active Share and Fund Duration: Four-Factor Alpha of Net Fund Returns

Table 4 reports the abnormal returns of the fund portfolios based on Active Share and Fund Duration. First, the funds are sorted in three classes of Active Share and then further sorted in three classes of Fund Duration, resulting in 9 portfolios. The first row of the table represents the alphas of the unconditional portfolios based only on Active Share and the first column the alphas of the unconditional portfolios only based on Fund Duration. The remaining nine alphas are the abnormal returns of the portfolios based on both Active Share and Fund Duration. Table 4 reports the equally-weighted Four-Factor alphas of the quarterly net fund returns in excess of the MSCI Europe small- cap index. The quarterly alphas are annualized. The significance levels of 10%, 5% and 1% are given by *, **, ***.

Equal Weighted Four-Factor Alphas of Net Fund Returns Active Share

Fund Duration Unconditional < 90% 90% to 95% > 95% Long-Short Unconditional -0.74 -1.00* -0.60 0.14 < 0.75 years -1.16 -0.45 -1.53** -1.16 -0.71 0.75 to 1.75 years -0.82** -0.82 -0.83* -0.73 0.1 > 1.75 years -0.21 -0.69** -1.45* 1.39 2.08** Long-Short 0.95 -0.24 0.08 2.55**

Among all portfolios, only high Active Share funds with patient investment strategies outperform, on average, their benchmark. These funds are in the top classes of both Active Share and Fund Duration and are defined as substantially different in stock holdings than their benchmark and have long

investment horizons. The outperformance of high Active Share and long Fund Duration funds is economically large but statistically not enough. As reported in table 4, these funds outperform, on

(32)

average, their benchmark by 1.39 basis points per year (T-Statistic of 1.62). This suggests that only managers with enough stock-picking skills to construct a substantially different portfolio than their benchmark and have strong conviction to hold these stocks for a longer period outperform their benchmark. Our result is consistent with the US mutual fund research of Cremers and Pareek (2016). They also found that only high Active Share funds with long Fund Duration, on average, show outperformance.

Funds with short investment horizons and low Active Share tend to underperform on average. Funds in the lowest classes of both Active Share and Fund Duration underperform their benchmark by -0.45 basis points (T-Statistic of 0.62). However, this result is economically small and statistically insignificant. Rather, we find that funds with short Fund Duration underperform in general, despite of Active Share class. The portfolio of funds with short Fund Duration and high Active Share underperform their benchmark considerably with an abnormal return of -1.16 basis points (T-Statistic of 0.92). This shows that only high Active Share funds with long investment horizons outperform their benchmark. Also, low Active Share funds, on average, underperform their benchmark. The lowest Active Share portfolios underperform their benchmark among all classes of Fund Duration. Even the portfolio

consisting of low Active Share and long Fund Duration funds underperforms with a significant abnormal return of -0.69 basis points (T-Statistic of 2.02). This shows that only patient funds with substantially different stock holdings are able to outperform their benchmark.

These results are consistent with the US mutual fund research of Cremers and Pareek (2016). However, the funds in our European small-cap sample show less underperformance than the sample of US mutual funds used in Cremers and Pareek (2016). For example, the lowest Active Share funds with shortest Fund Duration in the US sample have an abnormal return of -2.46 basis points compared to -0.45 basis points (T-Statistic of 0.62) in our European small-cap mutual funds sample. The reason for the better performance of our European small-cap sample could be that European funds, in general, perform better than their US counterparts, as documented by Otten and Schweitzer (2002). In addition, the small- cap premium can also be an explanation for the better performance. Our sample only consists of small-cap funds, whereas the US sample consists of small-, mid- and large-small-cap funds. Therefore, the small-small-cap premium could be the reason for the better performance of our sample.

Furthermore, our results show that high Active Share itself is not associated with outperformance. We find no economic and statistically significant evidence for high Active Share outperformance. Rather, our results show that high Active Share funds, on average, underperform their benchmark and only high Active Share funds with long Fund Duration outperform. The unconditional portfolio of high Active Share has an economically modest but statistically inadequate abnormal return of -0.60 basis points (T-Statistic of 1.21), as reported in Table 4. This result shows that Active Share is not associated with superior stock-picking skills, but rather a measurement of active management.

Moreover, we find no evidence that high Active Share funds outperform low Active Share funds in our sample. Across all Fund Duration classes, high Active Share classes show both out- and

(33)

underperformance relative to low Active Share classes and the differences are economically small and statistically insignificant. Only in the long Fund Duration portfolios, high Active Share funds outperform their low Active Share counterparts. Further, the unconditional high Active Share portfolio outperformed the unconditional low Active Share portfolio with an economically small abnormal return of 0.14 basis points (T-Statistic of 0.23). A possible explanation for this result could be that our sample only consists of small-cap funds and small-cap funds, in general, have high Active Share values. Consequently, the Active Share limits set in our research, 90% and 95%, could be too close together to report a significant

difference.

Also, patient investment strategies in itself are not associated with significant outperformance. Our results report that among portfolios with long Fund Duration, only funds with high Active Share outperform their benchmark. The unconditional long Fund Duration portfolio underperformed their benchmark by an economically small abnormal return of -0.21 basis points (T-Statistic of 0.26). This result suggests that patient investment strategies in itself are not associated with outperformance. Only patient funds with substantial different stocks holdings than their benchmark are able to outperform their benchmark.

However, long Fund Duration portfolios in our sample do outperform short Fund Duration portfolios. Among all Active Share classes, portfolios with long Fund Duration outperform most short Fund Duration portfolios. The unconditional long Fund Duration portfolio outperforms the unconditional short Fund Duration portfolio by an economically modest, but statistically not enough, abnormal return of 0.95 basis points (T-Statistic of 1.09). This result shows that patient investment funds, on average, perform better than frequently trading funds.

Active Share has the most impact on funds with patient investment strategies and the least impact on funds with short investment strategies. The long-short long Fund Duration portfolio, with long

positions in the long Fund Duration and high Active Share portfolio and short position in the long Fund Duration and low Active Share portfolio, has a large economic and statistically significant abnormal return of 2.08 basis points (T-Statistic of 2.24). While the abnormal returns of the long-short portfolios of funds with shorter Fund Duration are respectively, 0.1 Statistic of 0.13) and -0.71 basis points (T-Statistic of 0.67). This result suggests that high Active Share funds can add the most value with investing on long-term mispricing’s.

Likewise, long Fund Duration has the most impact on high Active Share funds. The long-short high Active Share portfolio has a large economic and statistically significant abnormal return of 2.55 basis points (T-Statistic of 1.94). While the abnormal returns of lower Active Share long-short portfolios are respectively 0.08 (T-Statistic of 0.12) and -0.24 basis points (T-Statistic of 0.31). This result suggests that holding stocks for a longer period add the most value when portfolios are substantially different then their benchmark.

For robustness, we also evaluate the European small-cap mutual fund performance with the one factor CAPM regression. Table 5 reports the annualized CAPM alphas of the equally-weighted portfolios.

(34)

The results confirm that only high Active Share and long Fund Duration funds outperform, on average, their benchmark.

The results of the CAPM regression are consistent with the Four-Factor model. The out- and underperformance of all the portfolios is similar in both regressions. In both regressions, only high Active Share funds with patient strategies outperform their benchmark. However, the CAPM alphas are slightly lower than the Four-Factor alphas. The portfolio with the highest Active Share and longest Fund Duration has, on average, an abnormal return of 1.18 basis point (T-Statistic of 1.44) compared to the 1.39 basis points (T-Statistic of 1.62) Four-Factor abnormal return. Also, the lowest Active Share and shortest Fund Duration portfolio has an abnormal return of -0.13 basis points (T-Statistic of 0.19) compared to the -0.45 basis points (T-Statistic of 0.62) Four-Factor abnormal return. The difference between the two regressions is due the exclusion of size, value and momentum factors in the CAPM regression.

Table 5. Active Share and Fund Duration: CAPM Alphas of Net Fund Returns

Table 5 reports the abnormal returns of the fund portfolios based on Active Share and Fund Duration. First, the funds are sorted in three classes of Active Share and then further sorted in three classes of Fund Duration, resulting in 9 portfolios. The first row of the table represents the alphas of the unconditional portfolios based only on Active Share and the first column the alphas of the unconditional portfolios only based on Fund Duration. The remaining nine alphas are the abnormal returns of the portfolios based on both Active Share and Fund Duration. Table 5 reports the equally-weighted CAPM alphas of the quarterly net fund returns in excess of the MSCI Europe small-cap index. The quarterly alphas are annualized. The significance levels of 10%, 5% and 1% are given by *, **, ***.

Equal Weighted CAPM Alphas of Net Fund Returns Active Share

Fund Duration Unconditional < 90% 90% to 95% > 95% Long-Short Unconditional -0.70 -0.90* -0.82* -0.13 < 0.75 years -0.81 -0.13 -1.06* -0.76 -0.62 0.75 to 1.75 years -0.82* -0.91 -0.89* -0.89* 0.02 > 1.75 years -0.37 -0.55*. -1.29* 1.18 1.73* Long-Short 0.43 -0.41 -0.23 1.94**

(35)

6. Conclusion

This research evaluates whether European small-cap mutual fund performance improves as Active Share and Fund Duration increase. We find that as Fund Duration increases, European small-cap mutual fund performance improves. However, among long Fund Duration funds, only high Active Share funds outperform their benchmark. We find no evidence for high Active Share outperformance in general. Only among long Fund Duration funds, high Active Share improves European small-cap mutual fund

performance.

Our results suggest that only managers with sufficient stock-picking skills to construct a

substantial different portfolio than their benchmark and have strong conviction to hold these stocks for a longer period are able to outperform their benchmark and add value for their investors. The average outperformance of funds with high Active Share and long Fund Duration is 1.39 basis points (T-statistic of 1.62). The average underperformance of funds with low Active Share and short Fund Duration is -0.45 basis points (T-statistic of 0.62). These results suggest that the European small-cap stock market is too information efficient in the short-term for fund managers to benefit from frequently trading. Conversely, the results suggest that managers with enough stock-picking skills to spot mispricing’s in the long-term are able to capitalize on the information inefficiencies in the long-term.

Our results show that longer Fund Duration improves European small-cap mutual fund performance. This suggests that patient strategies and holding stocks in a portfolio for a longer period increases European small-cap fund performance. An explanation for this result is that funds with patient strategies improve the performance of invested companies. The longer investment horizon set by asset managers reduces pressure on companies to maximize term financial results and eliminates short-termism at companies. Instead, the invested companies focus on long-term value creation and this increases company performance in the long-term and consequently increases fund performance.

We find less evidence for high Active Share outperformance in general. Only among long Fund Duration funds, high Active Share improves fund performance. This suggests that among European small-cap funds, the level of Active Share is not associated with fund performance, but rather a measurement of active management. An explanation for this results is that small-cap funds, in general, have high Active Share values. This suggest that the Active Share values in our sample are too close together to show performance difference between low- and high Active Share funds.

Moreover, the implementation of MIFID II in all likelihood reduces available research and increases information inefficiencies in the European small-cap market. Our results suggest that MIFID II does not create more investment opportunities for frequently trading European small-cap managers, because the European small-cap market is too information efficient in the short-term. However, in the long-term, MIFID II can add to the market inefficiencies and create more investment opportunities for small-cap managers with patient strategies to exploit.

European small-cap fund performance : outperformance of high active share funds with patient strategies