Is investor sentiment a missing factor for the explanation of mutual fund performance?

Universidade do Minho Escola de Economia e Gestão Tiago Almeida Ribeiro April 2022 Is investor sentiment a missing factor for the explanation of mutual fund performance? Tiago Almeida Ribeiro Is investor sentiment a missing factor for the explanation of mutual fund performance? UMinho|2022 Tiago Almeida Ribeiro April 2022 Is investor sentiment a missing factor for the explanation of mutual fund performance? Work developed under supervision of Manuel José da Rocha Armada Master's Dissertation Master in Finance Universidade do Minho Escola de Economia e Gestão ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição-NãoComercial-SemDerivações CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/ iii ACKNOWLEDGMENTS First, I would like to thank my supervisor, Professor Manuel José da Rocha Armada, because he was the one who supported me directly in this journey. I remember, from the very first meeting with him, that I was anxious about asking him to be my supervisor, since I didn’t know how our student/supervisor relationship would be. However, my passion for behavioral finance spoke louder and also the recognition of Professor’s Rocha Armada reputation as one of the best in the country in this area, made me to embrace this challenge. I have to say that it was one of the best decisions I could take in academia, because I have not only learned a lot with him, but also, I have acquired a good friend in my life. My supervisor has an extraordinary expertise in the area of finance, making him a unique professor and, alongside with that, I could understand, through our sometimes deep talks, that he is a fantastic person for those who prove to be worth its trust. So, I am really honored to conduct my investigation under his supervision. Secondly, I obviously acknowledge my entire family, especially my mother and my sister, which helped me every single second since I was born to be what I am today. Without them, it wouldn’t be possible to succeed. Their support is incommensurable, and I can only say, with these short text, that my gratitude is infinite. Thirdly, I also want to thank all of my friends (including my dogs), which were definitely the best I could ask for during the leisure times, in order to distract myself from the dissertation and gain additional energy to keep going with strength. Finally, I acknowledge Professor Miguel Portela as one of the determinant persons to enable me to perform this work. All my knowledge about Stata, that I have, is due to him, and for me learning a statistical software, without any base, is one of the most difficult parts for an inexperienced student starting its research. As you understand, I cannot mention everyone, but you know to whom I am referring to with these acknowledgments. Thank you all that are part of my life and give me motivation and purpose to get up every day and to strive for happiness. iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. v Será que o sentimento do investidor é um fator em falta para a explicação do desempenho dos fundos de investimento? RESUMO Este trabalho de investigação testa a inclusão da medida de sentimento do investidor AAII em modelos de finanças tradicionais, como o modelo de fator q-4 (Hou et al.,2015) e o modelo de 4 fatores de Carhart. Um dos principais objetivos consiste em avaliar se o sentimento do investidor pode representar uma das fontes do alfa dos fundos de investimento, que muitas vezes é visto como desempenho superior/inferior. Ao nível individual/fundo, a inclusão do sentimento do investidor reduz a probabilidade de ocorrência do alfa (quer seja positivo ou negativo), atribuindo um papel para essa variável comportamental no desempenho individual dos fundos de investimento. Uma validação de robustez, usando o índice composto de sentimento do investidor de Baker & Wurgler (2006), confirma os resultados, sobretudo para explicar o desempenho positivo. Ao nível das carteiras, os resultados mostram que o sentimento do investidor não é uma variável estatisticamente significativa para explicar as rendibilidades em excesso das carteiras vencedoras/perdedoras de fundos, com um nível de desempenho positive ou negativo associado, o que evidencia que o sentimento não representa um fator de prémio de risco nos designados modelos tradicionais. Em geral, o sentimento do investidor é um fator em falta a fim se de avaliar o desempenho dos fundos de investimento, uma vez que é uma das fontes do alfa. Por não ser considerado um fator de risco, também existe algo em falta nos modelos tradicionais de avaliação de ativos. As finanças comportamentais, ao considerar o sentimento do investidor, parecem ser resposta. Palavras-chave: Desempenho dos fundos de investimento; Finanças Comportamentais; Sentimento do investidor. vi Is investor sentiment a missing factor for the explanation of mutual fund performance? ABSTRACT This investigation tests the inclusion of the AAII investor sentiment measure into traditional finance models, such as: the q-4 factor model (Hou et al., 2015) and the Carhart 4-factor model. One of the main goals is to evaluate if the investor sentiment might be one of the sources of the mutual fund alphas, which is often seen as superior/inferior performance. At the individual/fund level, the inclusion of investor sentiment reduces the probability of occurrence of alpha (whether positive or negative), showing that it plays a role in mutual fund individual performance. A robustness check, using the Baker & Wurgler (2006) investor sentiment composite index, seems to confirm the findings, at least to explain positive performance. At the portfolio level, the results show that, investor sentiment, is not a statistically significant variable to explain winner/loser portfolios of funds’ excess returns, with an associated positive or negative level of performance, which provides evidence of sentiment not representing a risk premium factor in the so-called traditional models. Overall, investor sentiment is a missing factor in mutual fund performance since it is one of the sources of alpha. By not being considered as a risk factor, something is also missing in traditional asset pricing models. Behavioral finance, taking into account the investor sentiment, seems to be the answer. Keywords: Mutual fund performance; Behavioral Finance; Investor sentiment. vii TABLE OF CONTENTS 1.Introduction .................................................................................................................................. 1 2.Literature Review ........................................................................................................................ 3 2.1 Brief history of traditional performance measurement ................................................................. 3 2.2 Efficient market hypothesis and performance persistence ........................................................... 3 2.3 Behavioral Finance vs Modern Finance ...................................................................................... 4 2.4 How to measure investor sentiment? .......................................................................................... 5 2.4.1 Market data sentiment ........................................................................................................ 5 2.4.2 Text data sentiment ............................................................................................................ 7 2.4.3 Survey data sentiment ........................................................................................................ 7 2.5 Impacts of investor sentiment on securities returns .................................................................... 9 2.6 Mutual fund performance and investor sentiment .................................................................... 11 3.Research methodology............................................................................................................. 12 3.1 Why the AAII sentiment measure among others? ..................................................................... 12 3.2 The q-4 factor model .............................................................................................................. 12 3.3 Methodological steps to follow ................................................................................................. 13 4.Data ............................................................................................................................................ 16 4.1 Data sources and collection .................................................................................................... 16 4.2 Investor sentiment descriptive statistics ................................................................................... 17 5.Empirical results ....................................................................................................................... 19 5.1 Investor sentiment and mutual performance at the individual fund level ................................... 19 5.2 Robustness check................................................................................................................... 21 5.3 Winner and loser’s fund portfolio style analysis ........................................................................ 22 5.4 Inclusion of investor sentiment on the q-4 factor model at the portfolio level ............................ 25 5.4.1 Winner fund portfolio and bullish investor sentiment ......................................................... 25 5 However, there are other extreme cases where limits to arbitrage arise and make those opportunities unprofitable because there are transaction and taxes expenses involved or if profitable, not riskless (Shleifer & Vishny, 1997; Barberis & Thaler, 2005). There is, for instance, the risk of the ongoing mispricing worsening, in the short-term, when investors incorporate their biased beliefs and intuitions or simply non-relevant information to stock prices, causing noise traders risk, which can be very costly to arbitrageurs (see De Long et al, 1990). Theoretically, this gives rise to investor sentiment in financial markets. Empirically, investors’ sentiment was chosen as one of the most relevant behavioral factors to include in performance valuation models, since there is evidence which demonstrates that, sentiment, is a source of risk capable of explaining excess returns (Keiber & Samyschew, 2017; Keiber & Samyschew, 2019). 2.4 How to measure investor sentiment? One of the major literature concerns, regarding investor sentiment, consists of which methods to use in order to measure it. Secondly, the literature concentrates efforts on the evaluation of the respective impacts of sentiment on market returns. As to the first, there are indirect and direct measures of investor sentiment. Indirect measures are associated mainly with market variables, as it will be shown next, while direct ones correspond to survey indexes. Overall, investor sentiment measures can be classified into three broad categories: market, text, and survey data sentiment measures (Zhou, 2018). 2.4.1 Market data sentiment Ben-Rephael et al. (2012) uses market data, on mutual fund net flows, from stock funds to bond funds and vice-versa, depending on investors risk appetite, influenced by sentiment. In the previous study, there is evidence of high quantity of money inflows in equity mutual funds, when the market is going up, which translates into a short-term euphoria towards riskier investments, providing initial positive returns. Consistently, Ben-Rephael et al. (2019) also makes use of mutual fund net flows but, this time, to capture foreign sentiment, concluding that investors tend to overreact to non-local news (outside negativity bias), leading to shifts of money across international markets, depending on investors local conditions perceptions. Furthermore, the put/call ratio and the volatility index are two other ways of measuring investor sentiment by using market variables. The former represents the number of put options over the number call options traded in the market, while the latter analysis implied volatility by also using options to capture 6 sentiment. The underlying rationale is that, as they increase, both represent fear and pessimistic forecasts about the market, because a higher number of traded put options and a higher level of volatility are usually a signal of a bearish tendency. Reis and Pinho (2020) found that these two measures are strong predictors of sentiment and statistically significant for market returns. Despite that, Bandopadhyaya and Jones (2008) demonstrated that the put/call ratio is a better proxy for investor sentiment than the volatility index, since it captures variations non justified by economic factors more accurately. However, we have to take into account that the first paper uses an European S&P350 index, while the second uses the American S&P 500 index, so one cannot generalize the results, because a given measure could be a better predictor for a certain country or for a specific time period. An additional market measure, that has been used as a proxy for investor sentiment, is the closeend fund discount (CEFD), which represents the difference between the net asset value (NAV) and the market value of a close-end mutual fund. Lee et al. (1991) points out that CEFD accompanies variations in individual investor sentiment because, just like small stocks, they are more subject to be traded by non-professional investors, which are also the ones more prone to be “victims” of behavioral biases. In general, findings report that the lower is the fund discount in relation to NAV, the more bullish sentiment is, or that, there is a negative correlation between CEFD and investor sentiment (Lee et al., 1991; Neal & Wheatley, 1998). Notwithstanding, some researchers argue that CEFD is not a relevant factor to predict sources of systematic risk in asset pricing (Doukas & Milonas, 2004). Alternatively, Qiu and Welch (2004) found that CEFD is not even a suitable measure of investor sentiment, and so the literature has been abandoning this proxy gradually over time. Differently, Baker and Wurgler (2006) adopt an interesting method by forming a composite index (BW) with six market variables, acting as investor sentiment proxies: CEFD, market turnover, volume of IPO’s, first day returns on IPO’s, equity issues over total debt and equity issues, and the dividend premium. Firms have a higher probability of going public when investor sentiment is high in order to get more financing at a lower cost, which induces higher volume and first day of returns become very appetizing for investors, increasing general optimism (Ritter, 1991). Market turnover (or liquidity) is also considered as a proxy for investor sentiment in this framework 1 , since the higher the liquidity, the higher the investors’ mood, in a world with short-selling constraints, indicating a possible short-term overvaluation (Baker & Stein, 2004). 1 However, there are recent updates to the BW investor sentiment composite index. The market turnover variable was removed from the index, so now it only contains five market variables. 7 Baker and Wurgler (2007) complement their initial work, by concluding that there is an inverse relation between investor sentiment and subsequent stock returns and Baker and Wurgler (2006) already stated that, this effect, is higher for small, young, high volatility, non-dividend paying, unprofitable and distressed stocks which, in overall terms, are less liquid and so more likely to present costly and harder arbitrage opportunities to exploit. 2.4.2 Text data sentiment Now, regarding text data sentiment measures, Breitmayer et al. (2019) collected 14.9 million opinions from ShareWise social platform and found that stock opinions contain valuable information for expected future stock prices, implying a positive correlation with stock returns. No different conclusions were achieved by Renault (2017), which have chosen StockTwits social platform but applying the same methodology as the previous paper, for intra-day stock market returns. They just report, in addition, that after the initial positive returns, generated by a high level of sentiment, in the following trading day the reversal occurs, denoting evidence of noisy investors in the market. Duz Tan and Tas (2020) and Oliveira et al. (2017) adopt an alternative path by using Twitter as a social media that may unveil people’s general mood. For this second approach, positive and negative comments in Twitter are spotted through key words. The authors demonstrated a significant matching correlation between negative (positive) scores based on lexical characteristics and negative (positive) stock returns. McGurk et al. (2020) also demonstrates that positive and negative tweets are associated with investor sentiment, providing statistically significant abnormal returns for investors. Siganos et al. (2014) uses the Facebook Gross National Hapiness (FGNH) index, which has the same underlying idea as of the previous papers, since it separates positive and negative words in Facebook that signal a bullish or bearish investor sentiment. Once again, the results are similar, the index is able to predict daily returns that is followed by a short-term reversal during the next trading weeks. 2.4.3 Survey data sentiment Additionally, survey data sentiment measures, directly asking people about their prospects regarding market conditions, are also used as a proxy for investor sentiment. For instance, Fisher and Statman (2003) and Schmeling (2009) use an economic type of survey, called the Consumer Confidence Index (CCI), which represents consumption intensions and expectations for inflation and interest rates. It captures sentiment because people’s future perspectives are normally shaped according to the recent 8 experience and, unconsciously, people overestimate the probability for the occurrence of an event when it is still present in an individual memory. Moreover, economic perceptions are anticipated by financial market movements, since investors take into account the economic cycle when valuing a given market, because it contains important information to discount future anticipated cash flows. In that sense, Chung et al (2012) observed that investors’ optimism grows when the economy is expanding, which reflects a higher sentiment level. Inversely, when the economy contracts, investor sentiment tends to decrease, so it can also be seen as “regime dependent”. So, when consumers are optimist as to the economy, investor’s opinion, towards financial performance, in the next few months, are heavily influenced. One more relevant sentiment proxy is the American Association of Individual Investors (AAII) index, which consists of a financial type of survey. It has been widely used to analyze general stock market reuturns (Smales, 2017; He et al., 2019), specific industry stock returns (Sayim et al. 2013) and, even, to differentiate those impacts between distinct types of investors, according to their sophistication (Fisher & Statman, 2000). The AAII data describes three possible states, considering investors’ expectations for future market movements for the next 6 months, in a questionnaire format. They can be either bullish, bearish, or neutral, giving rise to three different investor sentiment proxies measured in percentual terms. Investor’s intelligence (II) survey sentiment index is, again, a financial survey used as a proxy to study the stock market and it basically has the same underlying logic as the AAII, with bullish, neutral, and bearish investors (Lee et al., 2002; Brown & Cliff, 2005). Kurov (2008) uses both II and AAII to investigate investor sentiment impacts on future markets, finding “positive feedback”, on trading trend. In addition, Chau et al. (2016) use a combination of sentiment proxies including the II and AAII measures, and they report that sentiment is linked to changes in risk tolerance and investors might take advantage of crowd euphoria/dysphoria by adopting contrarian trading strategies. Overall, market, text and survey data sentiment measures, have a significant predictive power for stock returns. Despite market and survey data being the most common implemented ones, text data measures are gaining more popularity recently and, in fact, there is a whole world to explore behind them, since the information provided on social media is vast, but it also requires more advanced programming techniques. For the purpose of this dissertation, the focus will be on the AAII survey data sentiment index, not only because of its simplicity, direct and relevant use and also, as it will be additionally explanained in the methodology section, in order to justify my choice. 9 2.5 Impacts of investor sentiment on securities returns When describing investor sentiment measures, I have been already discussing some of the associated impacts on returns. To complement the analysis, I will also mention the most relevant studies which better demonstrate how stock returns predictability, with sentiment, works. Firstly, Lee at al. (2002) start by providing empirical evidence that investor sentiment is a significant factor to explain equity excess returns and conditional volatility, suggesting a positive correlation between the excess returns and shifts in sentiment level. The magnitude of shifts in sentiment exhibits a significant impact on the formation of conditional volatility of returns and expected returns: changes in sentiment have a negative correlation with market volatility, since when investors become more bearish, volatility increases and vice-versa. These findings are in line with Chau et al. (2016), which shows the importance of sentiment and the idea that, trading activity is more notorious during bear markets, due to increased uncertainty. This leads to a magnified effect of bubbles when the stock market goes up through a positive feedback mechanism (see Shiller, 2002), and then crashes, induced by investors euphoria, followed by fear attitudes, which are represented by a massive selling pressure, and this kind of behavior based on investor sentiment increases the probability of a crisis (Zouaoui et al., 2011) Secondly, although using different econometric models, both Baker and Wurgler (2006) and Huang et al. (2015) find strong evidence that high (low) beginning of period sentiment, implies subsequent lower (higher) returns. They argue that “noise traders” try to benefit from early momentum bullish tendency, while arbitrageurs face huge costs in order to paddle against the market and correct ongoing mispricing. These results are consistent with Lux (1995), reveling a tendency for investors to overreact due to herding behavior. This leads to overoptimistic perception of future cash flows and, consequently, long-term reversals just like in Ben-Rephael et al. (2012), the initial good performance of equity mutual funds is reversed within ten months. Despite this fact, markets can remain inefficient before future reversals occur, because there are limits to arbitrage that make eventual opportunities unprofitable (Barberis & Thaler, 2005). There is also significant evidence that within the whole market, small cap and growth firms are the most impacted ones, by changes in sentiment, because they are difficult to arbitrage and so the stock price can remain different from its fundamental value for long periods of time (Baker & Wurgler, 2006; Baker & Wurgler, 2007; Smales, 2017). So, the inexistence of arbitrage opportunities does not mean that the market is efficient, nor that financial securities are fairly priced. Deviations from intrinsic value are, in fact, very common in financial markets, but sometimes not very pronounced and worthwhile 10 to be exploited. Nonetheless, the reported effects are equivalent to the previously described reversion effect. According to Smales (2017), investor sentiment has a statistically significant impact on stock returns and this relationship tends to be positive for contemporaneous returns and negative for future returns, providing further evidence of overvaluation followed by long-term corrections in stock prices. Additionally, Schmeling (2009) studied the effects of investor sentiment on international stock markets, concluding that it is a significant variable to predict stock returns for short periods of time. However, this result changes for some countries, in which investor sentiment does not have significant predictive power. At the country level, the impacts of sentiment on stock returns are higher when there is lower market integrity (for example, in a market with poor state and corporate governance), and lack of efficiency in regulatory institutions and for nations that are culturally more predisposed to herd-like behavior, which is consistent with the less liquidity and noise trading arguments, respectively. Similarly, Baker et al (2012) also study the global impacts of investor sentiment for different countries, stating that sentiment is a contrarian indicator of stock market returns, and that investor sentiment is contagious across markets. In contrast with the standard findings in the literature, Brown and Cliff (2004) show that there is an even stronger relationship between changes in sentiment and large cap stocks return. Moreover, they found predictive power for short-term returns to be quite small, which is also a different conclusion in comparison to previous studies. Conversely, Brown and Cliff (2005) update their conclusions, with new evidence, by finding that there is a positive contemporaneous stock price reaction to investor sentiment, but future returns are negatively correlated with sentiment, which denotes the reversal effect already discussed. Nevertheless, sometimes market inefficiencies tend to persist, especially for illiquid securities which require high transaction costs, in order to explore eventual arbitrage opportunities. Additional compensation for risk factors is an important feature to explain abnormal returns (Fama & French, 1993), but psychological biases also play a fundamental role (Daniel & Titman, 1999). It all depends on the timing of the analysis since financial markets are dynamic. Stock market prices might not be at their intrinsic value, in the short term, because of investor’s sentiment, but markets tend to become more efficient, in the long run, due to the fact that demand and supply are always acting towards the equilibrium fair value. So, it turns out that, modern finance and behavioral finance, might not be mutually exclusive, but rather complementary for the explanation of deviations from expected rates of return. In the research methodology section, insights will be provided, considering both perspectives. 11 2.6 Mutual fund performance and investor sentiment Until now, as previously discussed, the literature had an extensive focus on the effects of sentiment on stock market returns. To the best of my knowledge, there is lack of evidence about how investor sentiment influences mutual fund performance, since very few investigations were done based on that relationship. Ben-Rephael et al. (2012) uses mutual funds but to capture investor sentiment through net flows of money. As regards, exclusively, to the proposed topic, Bu (2020) is one of the pioneer studies in this emerging field by, precisely, studying the impacts of investor sentiment on mutual funds’ alpha. He concludes that both bullish, bearish, and bullish-bearish spread sentiment indexes, extracted by the AAII survey data, are statistically significant to explain the overperformance of mutual funds at the portfolio level, while not relevant at all for underperforming ones. Additionally, the author finds that the occurring probability of positive alphas diminishes when the bullish sentiment is included in the model and decreases, for negative alphas, when the bearish sentiment is added, providing a crucial role for investor sentiment as a power predictor of mutual funds’ performance. Bu and Forrest (2020) also reach the same conclusions by using, once more, the CRPS survivorship-bias-free fund data base. Moreover, they demonstrate that the AAII index presents better explanatory power in relation to the BW index. In a complementary form, Bu (2021a) further research on mutual fund alpha, provides evidence that the inclusion of sentiment in benchmark models is crucial to predict fund’s performance and the results are robust for sub-samples created for the entire period of study between 2000 and 2014, with only two notable exceptions: the 2007-2008 financial crisis and the fact that with the Fama and French 5-factor model the outperforming probabilities don’t change with the inclusion of the BW sentiment measure. As the author argued, this result might, in fact, be interpreted as a possible correlation between the selected investor sentiment measure and the risk factors incorporated in the model. Thus, due to a huge lack of research in this area, this research aims at filling a gap in the literature. Namely, investor sentiment and mutual fund performance has only been studied mainly by Bu, Q. and within the US context. So, this investigation, will be the very first one to conduct a study about investor sentiment and European mutual fund performance, where the selected funds’ portfolios invest mainly in US equity. 12 3. RESEARCH METHODOLOGY 3.1 Why the AAII sentiment measure among others? There are different ways of measuring investor sentiment, as I have discussed so far. The most recognized ones in the literature are: market, text, and survey data sentiment measures. Market sentiment measures, such as VIX (implied volatility of the S&P500) and bond yield spreads are, somehow, related to economic fundamentals, and that would possibly imply multicollinearity problems in the regression, due to the high correlation with the market factor. I also exclude the possibility of using text data sentiment measures because, although they can have a relevant use, most of the studies use them only to predict very short-term fluctuations in securities returns and those proxies are very time-consuming to construct, since it requires a lot of search and complex coding abilities to match lexical characteristics with sentiment. Because of this, I will choose the AAII sentiment index as a proxy for investor sentiment because this survey has the advantage of, effectively, capturing investors’ tendencies through one short and direct question. By measuring the % of bullish, neutral, and bearish investors, one can know the overall market expectations, about the near future in a simple and accurate way, without having concerns about different variables capturing identical phenomena. Moreover, Bu (2021b) shows that a direct measure of sentiment, like the AAII, is better to explain contemporaneous returns, while the BW indirect measure of sentiment has a lagged effect. 3.2 The q-4 factor model About the regression specification, many academics, doing reseach in mutual fund performance, have taken, for a long time period, the Carhart 4-factor model as the best model (Otten & Bams, 2002; Hunter et al., 2014; Andrikogiannopoulou & Papakonstantinou, 2019). Once again, I will challenge the previous approaches by, additionally, testing Hou et al. (2015) q-4 factor model, which includes a market factor, a size factor, an investment factor, and a profitability factor based on firms’ return on equity (Roe). The authors demonstrated that it is a better model in relation to Fama-French 6-factor model, through spanning tests, since the latter is subsumed by the former in terms of variables explanatory power. The model composition is as follows: 13 Image 1. Regression specification: Extracted from The Investment CAPM: Latest Developments - Research (theinvestmentcapm.com) Lu Zhang’s website. By implementing this new, and innovative basis model, also known as the Investment CAPM (Zhang, 2017), I am trying to conciliate both the supply and demand of financial markets as equally relevant determinants of returns. While the Investment CAPM takes into consideration a corporate finance point of view, for the explanation of the cost of capital, investor sentiment will be the demand side component to complement the logic of this framework. On the one hand, the intuition behind the supply perspective of the economy is that managers and their investment decisions, are critical for the company growth, because positive NPV projects are expected to generate higher profitability which will translate into lower future returns, as the market price increases. On the other hand, investors demand for financial instruments will also have an impact on returns, which could potentially be magnified by the fact that human beings are not fully rational, making aggregate sentiment an inevitable component on asset pricing. However, as Hou et al. (2015) and Zhang (2017) did demonstrate that, their model, better explains certain anomalies documented in the literature, I am aware of the possibility of investor sentiment becoming an irrelevant factor in the model, as sentiment tries precisely to accommodate for those anomalies. Nevertheless, as I have argued, sentiment can be a complement to account for mutual fund specific characteristics, rather than the whole equity market features, as often studied. Additionally, investor sentiment can also be a complement to explain mutual fund performance itself, outside the logic of risk premium factor interpretations that the so-called modern finance has introduced as a paragon. Besides that, I will also compare the results with the Carhart 4-factor model in order to have a more solid background of comparison. 3.3 Methodological steps to follow Until here, I have described both the theoretical model and the investor sentiment proxy intended to use as explanatory variables for mutual fund returns. Now the question is: what steps am I taking to achieve my goal? Just to clarify once more, the final objective is: to assess the importance of investor sentiment on mutual fund’s (abnormal) performance. 14 ❑ H1: Does the inclusion of investor sentiment reduces the probability of ocurrence of abnormal performance (e.g: in comparison with Carhart (1997), Hou et al. (2015) and Fama and French (2015) models?) So, after collecting the data, and regarding the first hypothesis, I use the q-4 factor model, the Carhart 4-factor model, the Fama-French 5-factor model and four different 5-factor sentiment based models (by adding the bullish sentiment and the bearish sentiment to the q-4 factor and Carhart 4-factor model), applying them individually to each mutual fund. The goal is to evaluate the occurring probability of alpha, under these models, to see their explanatory power in terms of arbitraging away the achieved alphas. Here, I expect that the probability of overperformance reduces significantly in the 5-factor bullish models, while the probability of underperformance is expected to decrease in the 5-factor bearish models, providing a role for investor sentiment on mutual funds’ performance. Secondly, for a robustness check, I will use Baker and Wurgler (2006) investor sentiment proxy, to assess if the occurring probability of alpha, also declines comparatively to the traditional finance models. ❑ H2: Is investor sentiment a statistically significant variable in order to explain excess returns of European mutual funds, which invest their portfolios in US equity, with an associated level of performance? Thirdly, and regarding the second hypothesis, I use the q-4 factor model to see which mutual funds presents positive and negative alphas over the sample period. Then, I rank those mutual funds into a “winners’ portfolio” separated from a “losers’ portfolio” according to their respective positive/negative performance measured by alpha, following Bu (2020) approach. Next, I will apply the chosen model for both portfolios in order to perform a short style analysis, since the characteristics of the categorized fund portfolios might contain fruitful information. Fourthly, I will add the sentiment proxy into the q-4 factor model and into the Carhart 4-factor model to test its overall significance level, on both winners’ and losers’ portfolios. As the AAII index provides three different states of investor sentiment (bullish, neutral, and bearish), I could get several combinations of regression outputs. Despite that, I will not use the neutral sentiment proxy, as it does not capture extremes on investor’s investment decisions and so the probability of having an impact on mutual 21 5.2 Robustness check In this section, I test the inclusion of the Baker and Wurgler (2006) investor sentiment composite index, in the Carhart 4-factor model and q-4 factor model, in a similar way and with identical conditions as in the previous section. The BW measure has been modified to contemplate only five economic variables (value-weighted dividend premium, first day IPO returns, IPO volume, close end fund discount and the equity shares in new issues), as market turnover has been dropped as one of the sentiment indicators. Although the BW is a market data sentiment measure, and so it might have some correlation problems with variables such as the book-to-market factor, which predicts similar economic issues, it is still one of the most accepted and well-known measures in the literature. So, I use it as a robustness check for the first hypothesis in this research. The results can be seen in the table below. Once again, the goal is to compare the so-called modern finance models with their sentiment-adjusted version, in order to see if the probability of occurrence of alpha is reduced by the inclusion of the BW investor sentiment measure. Table 3 Mutual fund performance and the BW index Occurring probability of alpha under the Carhart 4-factor model, q-4 factor model and the sentiment adjusted models where I add the BW investor sentiment measure. The results are based on monthly data from 31/12/2009 to 31/12/2018 due to shortage of data from 2018 onwards on the BW composite index. Statistical significance is considered at the 5% level with a tolerance margin of +0.5 p.p. of difference In the Carhart 4-factor model, 7,02% of the funds present a positive alpha, which represents a decrease of about 7 p.p. in relation to the result in table 2. This suggests that 2019 was a very good year to invest in financial markets, and, in fact, the S&P500 generated an almost 30% return in that year. Nevertheless, the argument I am trying to show here, is about the effect of sentiment on fund’s alphas. Thus, we can observe that the BW measure makes the probability of alpha to decrease until 3,86%, which seems to confirm the initial reported findings, with the AAII sentiment index. The main logic remains unchanged. On the one hand, it’s easier to earn abnormal returns when sentiment is high, so Alpha Carhart 4factor model q-4 factor model Carhart 4factor model plus BW q-4 factor model plus BW Positive 7,02% 10,18% 3,86% 6,67% Negative 4,91% 5,61% 4,91% 5,61% 22 when the BW value reaches relatively higher levels, compared to its average, a higher performance benchmark is set. On the other hand, the probability of a loss, in the market, is higher when sentiment is low, which sets a lower performance benchmark. However, this time, the underperforming probability remains unchanged in both models, suggesting that the BW measure is somehow limited to be the source of underperformance. Taking into account this fact, I can also mention that, this particular result is relatively similar to Hou et al. (2015), since despite their BW-adjusted models being able to reduce the probability of underperformance, the magnitude of that change is only, on average, 0,58 p.p. When I use the q-4 factor model, no different conclusions are reached. About 10,18% of the funds outperform the market in the q-4 model, while in the BW sentiment adjusted q-4 factor model, where the BW sentiment proxy is added to the q-4 factor model, only 6,67% of the funds beat the market, which also represents a considerable reduction. These robustness checks confirm that investor sentiment is a missing factor in European mutual fund performance evaluation at the individual level, since this behavioral variable is capable of predicting, at least, positive alphas. 5.3 Winner and loser fund portfolio style analysis Now, regarding the second research question, once again, first I apply the q-4factor model to each mutual fund in order to get the intercept value of the regression, which is the abnormal return. A positive constant means that the fund has outperformed the market index, whereas a negative constant signs underperformance. However, I only consider intercept values statistically significant at the 5% level 3 . In total, I get 46 winner funds and 20 loser funds, which are used to form an equally weighted winner fund portfolio and loser fund portfolio, respectively. Then, once more, based on the q-4 factor model, I perform a style analysis of these portfolios, in order to gather information about their composition. The results are as follows: 3 There are three loser funds which are only statistically significant at the 8% level, but I decided to include them due to shortage of data to form a robust portfolio. 23 Table 4 Winner fund portfolio Regression output for the winner fund portfolio based on the q-4 factor model. R_MKT, R_ME, R_IA and R_ROE stand for the market, size, investment, and profitability factors. The results are based on monthly data from 31/12/2009 to 31/12/2019 Statistical significance is considered at a 5% level. Table 5 Loser fund portfolio Regression output for the loser fund portfolio based on the q-4 factor model. R_MKT, R_ME, R_IA and R_ROE stand for the market, size, investment, and profitability factors. The results are based on monthly data from 31/12/2009 to 31/12/2019. Statistical significance is considered at a 5% level. As presented above, Tables 4 and 5 show that the market loading is positive and statistically significant for both portfolios. Nonetheless, the loser fund portfolio has a market excess return loading of about 0.90, which is above the 0.61 value for the winner portfolio. This means that, in terms of aggregated holdings, the loser portfolio is more approximated to the market portfolio, which in traditional theory has a coefficient equal to 1. _cons .0056415 .0023154 2.44 0.016 .0010551 .0102279 R_ROE -.0198776 .1304661 -0.15 0.879 -.2783059 .2385507 R_IA -.0643296 .1423342 -0.45 0.652 -.3462662 .217607 R_ME .151005 .1066534 1.42 0.160 -.0602549 .3622648 R_MKT .6105325 .0661899 9.22 0.000 .479423 .741642 RiR_F Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .136599627 119 .001147896 Root MSE = .02352 Adj R-squared = 0.5181 Residual .063618068 115 .000553201 R-squared = 0.5343 Model .072981559 4 .01824539 Prob > F = 0.0000 F(4, 115) = 32.98 Source SS df MS Number of obs = 120 _cons -.0034583 .0005954 -5.81 0.000 -.0046376 -.002279 R_ROE .0007919 .0335469 0.02 0.981 -.065658 .0672418 R_IA .0259116 .0365985 0.71 0.480 -.046583 .0984062 R_ME -.1006634 .0274239 -3.67 0.000 -.1549849 -.046342 R_MKT .9028947 .0170195 53.05 0.000 .8691823 .936607 RiR_F Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .131901642 119 .001108417 Root MSE = .00605 Adj R-squared = 0.9670 Residual .004206192 115 .000036576 R-squared = 0.9681 Model .127695451 4 .031923863 Prob > F = 0.0000 F(4, 115) = 872.82 Source SS df MS Number of obs = 120 24 In addition, and as expected, the constant term is positive (negative) and statistically significant for the winner (loser) portfolio, which reflects the positive (negative) performance of each individual fund comprising the portfolios. Regarding the remaining three factors (size, investment, and profitability), I can observe that, by comparing the factor loadings on both portfolios, they have contrarian signs: the size factor is positive for the winner fund and negative for the loser fund, whereas the investment and profitability factors are negative for the winner fund and positive for the loser fund. These results also reflect the style of the funds representing the two constructed portfolios. For example, the size factor in the loser portfolio is negative and statistically significant, meaning that those funds invest mainly in large capitalization stocks. Theoretically, the positive size factor for the winner portfolio means that the portfolio is composed by mutual funds that invest a significative proportion of their holdings in small capitalization stocks. However, I went to check the actual style of some of the mutual funds in both portfolios. I conclude that the loser portfolio is indeed composed by large capitalization firms, and the same applies for the winner portfolio, which confirms the statistical insignificance of the winner portfolio for this particular factor. Additionally, a positive investment factor loading means that the portfolio is composed by firms with a lower investment/total assets ratio, which is the case of value firms, and, in fact, this factor is negatively correlated with the book-to-market factor used in traditional models (see Hou et al., 2015). Although the investment factor loadings are not statistically significant for both models, the analysis of each fund allows me to conclude that the previous sentence is correct. The loser fund portfolio, presenting a positive investment factor, is mainly composed by value firms, while the winner portfolio, which has a negative investment factor value, has a clear tendency to invest in growth funds. In this latter case, the value is negative since growth firms, at least in the last decade, are perceived to be less risky than value firms, meaning that investors demand a relatively lower risk premium. Lastly the profitability factor, measured by the firm’s ROE also presents consistent results with what would be expected. The winner portfolio shows a negative coefficient for this factor, since the portfolio is composed by profitable firms (who are also the ones which outperform the market index), so a lower risk premium is required. On the other hand, the positive ROE factor for the loser portfolio means that the loser funds invest in lower profitable stocks, and so investors require a higher risk premium to compensate them. 25 5.4 Inclusion of investor sentiment on the q-4 factor model at the portfolio level In this section, I test whether the inclusion of investor sentiment is relevant at the performance portfolio level, in an asset pricing perspective, to answer the second research question. To do that, I add the bullish sentiment index to the q-4 factor model in order to observe its impact on the winner fund portfolio. The same procedure is adopted to the analysis of the loser fund portfolio, but this time with the inclusion of the bearish sentiment index instead. The results can be consulted below in tables 6 and 7. 5.4.1 Winner fund portfolio and bullish investor sentiment Table 6 Winner fund and AAII bullish sentiment index I Regression output for the winner fund portfolio based on the bullish sentiment-adjusted q-4 factor model. R_MKT, R_ME, R_IA, R_ROE and Bullish stand for the market, size, investment, profitability and positive investor sentiment factors. The test results are based on monthly data from 31/12/2009 to 31/12/2019. Statistical significance is considered at a 5% level. _cons -.001821 .0111518 -0.16 0.871 -.0239127 .0202707 Bullish .0212821 .0311071 0.68 0.495 -.0403409 .082905 R_ROE -.0271598 .1312014 -0.21 0.836 -.2870687 .2327492 R_IA -.0764121 .1437535 -0.53 0.596 -.3611866 .2083624 R_ME .1380831 .1085566 1.27 0.206 -.0769666 .3531329 R_MKT .5954288 .0699202 8.52 0.000 .4569174 .7339402 RiR_F Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .136599627 119 .001147896 Root MSE = .02357 Adj R-squared = 0.5158 Residual .06335793 114 .000555771 R-squared = 0.5362 Model .073241697 5 .014648339 Prob > F = 0.0000 F(5, 114) = 26.36 Source SS df MS Number of obs = 120 26 5.4.2 Loser fund portfolio and bearish investor sentiment Table 7 Loser fund and AAII bearish sentiment index I Regression output for the loser fund portfolio based on the bearish sentiment-adjusted q-4 factor model. R_MKT, R_ME, R_IA, R_ROE and Bearish stand for the market, size, investment, profitability and negative investor sentiment factors. The test results are based on monthly data from 31/12/2009 to 31/12/2019. Statistical significance is considered at a 5% level. The underlying intuition to this methodology is that bullish sentiment is a possible factor to explain eventual relatively higher generated excess returns due to market euphoria, while the opposite applies to bearish sentiment. I can conclude that, although bullish(bearish) sentiment has a positive (negative) coefficient, the results aren’t statistically significant. So, it seems that including investor sentiment isn’t relevant at all, in the so-called traditional models. However, I cannot immediately conclude that investor sentiment is not considered a risk factor. As I remember from the methodology section, the construction of the factors in the q-4 model is designed in a way to explain many anomalies detected in the literature, such as the value-growth anomaly and momentum (Hou et al., 2015). Investor sentiment tries precisely to explain some of those anomalies, in a behavioral perspective, but as Zhang (2017) as demonstrated, the anomalies become regularities in the q-4 factor model, because the investment and the profitability factors are able to capture most of them. _cons -.000436 .0028875 -0.15 0.880 -.0061562 .0052841 Bearish -.009588 .0089639 -1.07 0.287 -.0273454 .0081694 R_ROE .0011669 .0335277 0.03 0.972 -.0652512 .0675851 R_IA .0285067 .036656 0.78 0.438 -.0441085 .101122 R_ME -.1045992 .0276526 -3.78 0.000 -.1593789 -.0498196 R_MKT .8938 .0190157 47.00 0.000 .8561301 .9314699 RiR_F Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .131901642 119 .001108417 Root MSE = .00604 Adj R-squared = 0.9670 Residual .004164398 114 .00003653 R-squared = 0.9684 Model .127737244 5 .025547449 Prob > F = 0.0000 F(5, 114) = 699.36 Source SS df MS Number of obs = 120 27 5.5 Inclusion of investor sentiment on the Carhart 4-factor model at the portfolio level Thus, I cannot necessarily conclude straight-away that something is missing in the so-called traditional models because I am using a different from usual regression specification, by implementing the Hou et al. (2015) model. Probably investor sentiment is simply not a statistically significant variable in this q-4 model in terms of asset pricing at an aggregated level due to the model specification. The robustness of the q-4 factor model is so high that, the composing factors can mostly summarize the cross-section of market returns, as the authors of the model have argued. In order to gain additional comprehension about this issue, I examine the inclusion of investor sentiment in the Carhart-4 factor model, which is a more standard and tested approach in the literature. The results can be seen below in tables 8 and 9. 5.5.1 Winner fund portfolio and bullish investor sentiment Table 8 Winner fund and AAII bullish sentiment index II Regression output for the winner fund portfolio based on the bullish sentiment-adjusted Carhart 4-factor model. R_MKT, SMB. HML, Mom and Bullish stand for the market, size, book-to-market, momentum and positive investor sentiment factors. The test results are based on monthly data from 31/12/2009 to 31/12/2019. Statistical significance is considered at a 5% level. _cons -.0013379 .010867 -0.12 0.902 -.0228654 .0201897 Bullish .0179927 .0302043 0.60 0.553 -.0418418 .0778272 Mom .1231536 .0747715 1.65 0.102 -.0249683 .2712754 HML -.0182608 .1032524 -0.18 0.860 -.222803 .1862814 SMB .1752698 .1033278 1.70 0.093 -.0294219 .3799614 R_MKT .6221252 .0660472 9.42 0.000 .4912862 .7529643 RiR_F Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .136599627 119 .001147896 Root MSE = .02317 Adj R-squared = 0.5322 Residual .06121513 114 .000536975 R-squared = 0.5519 Model .075384498 5 .0150769 Prob > F = 0.0000 F(5, 114) = 28.08 Source SS df MS Number of obs = 120 28 5.5.2 Loser fund portfolio and bearish investor sentiment Table 9 Loser fund and AAII bearish sentiment index II Regression output for the loser fund portfolio based on the bearish sentiment-adjusted Carhart 4-factor model. R_MKT, SMB. HML, Mom and Bearish stand for the market, size, book-to-market, momentum and negative investor sentiment factors. The test results are based on monthly data from 31/12/2009 to 31/12/2019. Statistical significance is considered at a 5% level. Table 8 and 9 show that the bullish and bearish sentiment indexes present statistically insignificant coefficients. Since the Carhart 4-factor model is a general accepted model in finance theory and has similar foundations in relation to the other well-known modern finance models, I can now state that something is really missing in asset pricing theory. By confronting the two-research hypothesis under which I conducted my investigation, I can observe that investor sentiment is one of the sources of fund performance, but, because it does not have a role as a risk premium factor, the so-called traditional models cannot capture the importance of investor sentiment in mutual fund performance. _cons -.000526 .0027852 -0.19 0.851 -.0060435 .0049916 Bearish -.0086372 .0086488 -1.00 0.320 -.0257704 .0084961 Mom -.0221568 .0188595 -1.17 0.243 -.0595174 .0152038 HML .0297794 .0260499 1.14 0.255 -.0218253 .0813842 SMB -.1185083 .0259095 -4.57 0.000 -.1698348 -.0671818 R_MKT .8861659 .0179237 49.44 0.000 .8506591 .9216727 RiR_F Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .131901642 119 .001108417 Root MSE = .00584 Adj R-squared = 0.9692 Residual .003893819 114 .000034156 R-squared = 0.9705 Model .128007823 5 .025601565 Prob > F = 0.0000 F(5, 114) = 749.54 Source SS df MS Number of obs = 120 29 6. CONCLUSIONS Returning to the starting point, it goes again without saying that many investment professionals rely on the so-called traditional finance models to assess the ability of mutual fund managers to earn abnormal returns. The problem is that both the market and funds excess returns are dependable on the economic cycle, which could be favorable or not. Risk premium factors are a well-established source of explanation to the fluctuations in different financial markets cycles. However, not all those variations can be explained by risk-based rational interpretations. And behavioral finance has showed that asset pricing mistakes can be made when investors incorporate their non-rational based expectations into securities prices, giving raise to what we call behavioral anomalies. Thus, to conduct this research, we believed since the beginning, that something is missing in asset pricing. Specifically, a behavioral variable which can potentially be capable of improving the financial models, by accounting for investors irrationality in the decision-making process, affecting demand and supply. The AAII Investor sentiment was the chosen variable to eventually play a role because it is a dynamic proxy of the economic cycle. Investor’s bullishness and bearishness gives information about the actual state of a financial market. And, in fact, since the economic cycle is a crucial component of the performance for any security, investor sentiment came as a possible missing factor to be the source of alpha. So, this investigation aimed to study mutual fund performance by testing the inclusion of investor sentiment in benchmark models. Namely, the bullish and bearish investor sentiment proxies were tested at two different levels: individual/fund and portfolio level. On the one hand, the bullish sentiment adjusted q-4 factor and Carhart 4-factor model reduce the probability of outperformance, in comparison to their traditional peers, at the individual fund level. On the other hand, the bearish sentiment adjusted models reduce the probability of occurrence of negative performance in comparison to the standard q-4 and Carhart models. The intuition is that bullish sentiment sets a higher benchmark for mutual fund managers to earn positive alpha, while bearish sentiment sets a lower benchmark for managers to underperform. This means that it’s easier to beat the market when sentiment is high and easier to underperform when sentiment is low. So, the source of fund alpha, at the individual level, comes, but not exclusively, from investor sentiment, providing this behavioral variable a role into the explanation of European mutual performance, which invest their holdings mainly in US equity. 30 A robustness check using one additional investor sentiment measure (Baker & Wurgler, 2006) for this first research question confirms these results, at least for the explanation of positive performance. The probability of earning positive alpha diminishes with the inclusion of the BW sentiment measure, while it maintains equal as to the negative performing funds. Regarding the second hypothesis, a winner fund and loser fund portfolios were constructed based on the q-4factor model, in order to test investor sentiment in a joint asset pricing perspective. Firstly, the style analysis results reveal that the loser funds stock exposure is similar to the market portfolio benchmark, since both invest primarily in large capitalization and value firms. So, one of the implications of the results, is that to be in the winner funds category, mutual fund managers need to deviate from the market portfolio, by investing, for example, in smaller and growth firms, with potential profitable opportunities in the horizon. Secondly, and more importantly, the investor sentiment proxy was included, into the q-4 factor model, to analyze the effects on the constructed portfolios. The bullish and bearish indexes are not statistically significant to explain the winner and loser fund portfolio returns when using the q-4 factor model. However, one possible explanation could be that, the q-4 factor model is precisely designed in a way, to accommodate for many anomalies documented in the literature. The use of the profitability and investment factors is proven to be an effective way of summarizing the cross-section of securities returns (see, Horstmeyer et al., 2022, January 10), which gives Hou et al. (2015) a relevant recognition, in relation to the robustness of the q-4 model. Because of that, the inclusion of the same investor sentiment proxy, on the Carhart 4-factor model, was tested. Once more, the bullish and bearish sentiment proxies do not show statistically significant coefficients. So, is investor sentiment a missing factor for the explanation of European mutual funds’ performance, which invest mainly in US equity? At the individual fund level, it seems that investor sentiment is, indeed, a missing factor. And, at the portfolio level, the answer is complementary. From the first research question, we can conclude that investor sentiment is one of the sources of European fund performance. Moreover, from the second research question, one concludes that investor sentiment is not a risk premium factor. So, one might ask why, in the second research question, the so-called traditional asset pricing models do not recognize the importance of investor sentiment, when, in fact, that importance exists, taking into account the first research question? And the answer is that, not only investor sentiment is a missing factor, in mutual fund individual performance, but also that, the traditional finance models fail to capture the effects of sentiment, in an asset pricing perspective, since their factors are constructed to be risk premium interpretations of excess returns. But, as this investigation suggests, what is really 37 Appendix 1 –Alpha estimates for the first research hypothesis by model Mutual fund name Carhart 4factor model Fama-French 5factor model Bullish Carhart 4-factor model Bearish Carhart 4-factor model q-4 factor model Bullish q-4 factor model Bearish q-4 factor model 3 BANKEN AMERIKA .00374 .00454 -.00945 .02788* .00458 -.00979 .02864* ACOMEA AMERICA A1 .00034 -.00056 -.0022 .00533 .00027 -.00339 .00875 ALLIANZ AZIONI AMERICA .00142 .00182 -1.00e-05 .01484 .00175 -.00049 .01523 ALLIANZ RCM US EQUITY A .00226 .00248 -.01081 .02558* .00264 -.01184 .02634* AMERI GAN D GROUPAMA .00363 .00438* -.00874 .02058* .0039 -.00785 .02102* AMONIS EQUITY US CANADA .00565** .00715*** .00072 .01609 .00692*** .00104 .01453 AMUNDI AKTIEN ROHSTOFFE A -.00595 -.00822** -.01237 .01978 -.00569 -.01581 .02657 AMUNDI AZIONARIO .00127 .00198 -.00514 .0187** .00152 -.00535 .01826* ANIMA AMERICA A .00402* .00441* -.00355 .01539 .00413* -.00319 .01565 ARNIKA ARVEST AM.STARS .00378 .00285 .00797 .00945 .00355 .00752 .01353 ASI AMER.UCND.EQ. RET .00284 .00372 -.00339 .02135 .0037 -.00402 .02139 ASI AMERICAN EQUITY .00277 .00289 -.00604 .01618 .00272 -.00551 .01723 ASI STANDARD LIFE NORTH AMERICA .00346 .00403 -.00705 .0217 .00405 -.00699 .02222 AVIVA AMERIQUE I .00334 .00383 .00047 .01333 .00375 .00084 .01341 AXA FRAMLINGTON GLOBAL TECHNOLO .00418 .00621* -.01381 .03689** .00643* -.01899 .03325* AXA INDICE USA A C .00484** .0052** -.00133 .01459 .00483** -.00024 .01534 AXA ROSENBERG AMERICAN .00402 .00425 -.00934 .0202 .00409 -.00876 .0209 BANKIA BOLSA USA .00074 .00154 -.01007 .0171 .00091 -.00853 .01741 BANKIA INDICE S&P 500 -.00196*** -.00205*** -.00197 -.00085 -.00226*** -.00176 -.00065 BATI ENTREPRENDRE .00272 .00336 -.01034 .02944** .00333 -.01185 .02995** BBVA BOLSA USA .00226 .00267 -.00315 .00792 .00257 -.00294 .00856 BBVA INDICE USA PLUS -.00421*** -.00423*** -.00579** -.0046** -.00414*** -.00633*** -.00467** BCV ENHANCED US -.00156*** -.00152*** .00009 -.00126 -.00163*** .00021 -.00124 BLACKROCK US DYNAMIC INC .00345 .00397 -.01225 .02414* .00319 -.0107 .02475* BLACKROCK US OPPS.AC .00294 .00354 -.00889 .0237* .00348 -.00874 .02522* BLLE.GIFF.AMER.A AC. .0053 .00677** -.00255 .02811* .00778** -.00755 .02559 BMO NORTH AMERICAN EQUITY 1 .00407 .00416 -.0109 .02052 .00386 -.00954 .02171 BMO US SMALLER COMPANIES .00464 .00467 -.00678 .02032 .00473 -.00554 .02288 BNL AZIONI AMERICA .00226 .00314 -.00554 .0211* .00315 -.00679 .01925 BNY MELLON US OPPORTUNITIES .00087 .00238 -.01176 .02423 .00228 -.01459 .0238 BSO AMER(C) SEGIPA .00115 .00143 .00032 .00505 .00168 -.00059 .00508 C-QUADRAT GLOBAL EQUITY ESG R - .00164 .00274 -.01776 .01825 .00131 -.01532 .01751 CAIXA ACOES EUA .00318 .00377 -.00672 .01635 .00345 -.00634 .01734 CAIXABANK BOL SELECCION USA .00215 .0027 -.00583 .01294 .00249 -.00546 .01371 CAIXABANK BOLSA USA .00311 .00335 -.00174 .01135 .00304 -.00055 .0123 CAJA INGENIEROS BOLSA .00193 .00213 -.00085 .00787 .00174 .00042 .00823 CD AMERIQUE STGIS.A .0027 .00338 .00217 .01425 .00313 .00307 .01498 CLEMED.INV.IOM US EQ .00436 .00458 -.01142 .02362* .0043 -.0104 .02478* CM-AM INDICIEL AMERIQUE 500 C .00433* .00484** -.00166 .01508 .00448* -.00072 .0158 COLUMBUS AMERICAN -.00191 -.00166 -.01345* .00541 -.00194 -.01123 .00575 COLUMBUS US MARKET -.0042* -.00396 -.00864 -.00361 -.00398 -.00626 -.00235 COVEA ACTIONS AMERIQUE A .00334 .00383 .00047 .01333 .00375 .00084 .01341 CPR USA ESG - P (C) .00413* .00453* -.00374 .01675 .00407* -.0025 .01765 DANSKE INVEST BIOTEKNOLOGI .00674 .00935* -.00589 .03733 .01224** -.01535 .03589 DANSKE INVEST TEKNOLOGI .00512 .00764** .00351 .02233 .00754** -.00106 .0183 DANSKE INVEST USA -.00238** -.00239** -.00254 .00177 -.00257*** -.00198 .00309 DANSKE INVEST USA DKK D .00416 .00454* -.00265 .01716 .00412 -.00132 .01908 DIT BIOTECHNOLOGIE .00438 .0065 -.00025 .01662 .01032* -.01039 .01679 DPAM CAPITAL .00504** .00546** -.00084 .01593 .00517** -.00007 .0167 DPAM CAPITAL B .00598** .00605** .00134 .0126 .00553** .0034 .01382 DPAM CAPITAL B EQUITIES US .00313 .00348 -.00395 .01296 .00285 -.00295 .0145 DPAM DBI-RDT B .00472 .00482 -.01698 .03334** .00483 -.01669 .03539** DWS INV.NORDAMERIKA .00383 .00472* -.00891 .03302*** .00533** -.01184 .03226** EDMOND DE ROTHSCHILD .00376 .00495** .00236 .01499 .00493** .0013 .01374 EOLE (C) .00368 .00478* -.00417 .01521 .00464* -.00403 .01462 ERSTE BEST OF AMERICA .00321 .0033 -.01672 .03411** .00384 -.01848 .03557*** ESPA STOCK AMERICA .00224 .00259 -.00862 .02528* .0026 -.00764 .02666** ESSOR USA OPPORTUNITIES P .00416 .00496* -.00208 .01934 .00584** -.00448 .01874 ETOILE ACTIONS US .00461** .00504** .00011 .01253 .00482** .00078 .01306 ETOILE MULTI GESTION .00228 .00294 .00021 .01163 .0027 -.00033 .01183 ETOILE MULTI GESTION ETATS-UNIS -.00307*** -.0031*** -.00038 -.00305 -.00314*** -.00099 -.00305 EURIZON AM AZIONI .00065 .00104 -.00862 .02039* .00067 -.00933 .02109* EURIZON AZIONI AMERICA .00277 .0034 -.00134 .01424 .00305 -.00105 .0151 FEDERAL INDICIEL US .00473** .00514** -.00222 .01746 .00483** -.00106 .01861* FID.AM.INSTL.FD. .00406 .00443* -.01321 .02708** .00404 -.01207 .02825** FIDELITY AMERICAN A ACC .00251 .00334 -.01349 .02151 .0032 -.01286 .02248* FIDELITY AMERICAN SPECIAL .00393 .00405 -.01402 .02388* .00358 -.01172 .02609* FIDEURAM MASTER SEL.EQ .00203 .00261 -.01004 .02082* .0024 -.01032 .02104* Panel A: European equity mutual fund's alpha 38 Appendix 1 –Alpha estimates for the first research hypothesis by model (continued) FONDMAPFRE BOLSA AMERICA .00303 .00317 -.00258 .01165 .00228 -.00024 .01251 FOURPOINTS AMERICA .00226 .00248 -.00131 .00628 .0025 -.00084 .00745 GAM NORTH AMERICAN GROWTH A .00564* .00551* .00111 .00908 .00542* .00306 .01109 GROUPAMA US EQUITIES .00391 .00467* -.00681 .01999* .00405* -.0063 .02026* GUT.KAPITALANLAGE .00214 .00241 -.00018 .02076 .0026 -.00033 .02239 GUTZWILLER FON.MAN.ONE .00203 .00257 -.00059 .0126 .00246 -.00009 .01445 HALIFAX NORTH AMERICAN C .00451 .00461 -.01064 .02187 .0044 -.00963 .02299 HANDELSINVEST NORDAMERIKA .0045* .00547** -.00273 .0264** .00506* -.00381 .02583* HSBC AMERICAN INDEX .00518* .00534* -.00718 .02258* .00524* -.00627 .0236* IBERCAJA BOLSA USA .00292 .00293 -.00026 .01065 .00248 .00042 .01192 ING DIRECT FONDO NARANJA .00438* .00483* .0003 .01294 .00443* .00068 .01314 INVESCO US EQUITY (UK) .00263 .0023 -.00332 .01734 .00268 -.00426 .02009 INVESTITORI AMERICA .0036* .00378* -.00159 .01343 .00349* -.0004 .01445 IQAM QUALITY EQUITY .00475* .00475* -.00964 .01849 .0043 -.00715 .01908 JANUS HENDERSON INST .00417 .00435 -.00506 .0196 .00438 -.00483 .02079 JANUS HENDERSON US GROWTH A .0025 .00334 -.01726 .02326 .00372 -.01869 .02283 JPM US EQUITY .00615** .00597** -.00631 .0177 .00567* -.00388 .02006 JPM US SELECT .00398 .00444 -.01104 .02567* .00418 -.01112 .02621* JPM US SMALL CAP GROWTH A .00394 .00555 -.00711 .02371 .00686* -.01132 .02189 JUPITER NTH AMERICAN INC .00403 .00409 -.00595 .01733 .00381 -.00426 .01965 KATHREIN US-EQUITY .00302 .00365 -.01377 .0307** .00404 -.01417 .03124** KBC EQ.FD.BUYBACK AM .00437* .00465* -.00296 .01415 .00437* -.00117 .01547 KBC EQ.FD.SMALL CAPS .00454* .00464* -.00185 .01136 .00433* .00052 .01279 KBC EQ.FD.US VALUE CAP .00488** .00512** -.00005 .01334 .00454* .00218 .0151 KBC EQUITY NORTH AMERICA .00397 .00463* .00003 .01261 .00423* .00115 .01338 KBC INSTL.FD.US EQ. CAP .00518** .00588** .00144 .01284 .0055** .0022 .01365 KEPLER US AKTIENFONDS A .0051* .00522* -.01765 .02976** .0047 -.01658 .03078** KLP AKSJE USA INDEKS .00508** .0055** -.00209 .01677 .00523** -.00125 .01772 KLP AKSJE USA INDEKS II -.00337* -.00351** -.00386 -.00043 -.00285* -.00488 .00246 KUTXABANK BOLSA EEUU -.00222*** -.00266*** -.01058*** -.00084 -.00306*** -.00915** -.00014 LABORAL KUTXA BOLSA -.00368*** -.00331*** -.00837* .00333 -.00372*** -.00815* .00261 LANSFORSAKRINGAR .00521** .00559** -.00757 .0217* .00528** -.00619 .0227* LANSFORSAKRINGAR USA .00425 .00466* -.00751 .02184* .00448 -.00771 .02247* LAZARD ACTIONS .00255 .0032 -.00222 .01506 .00269 -.00066 .0161 LBPAM ISR ACTIONS AMERIQUE .00243 .0031 -.00432 .01669 .00289 -.00399 .01795 LIONTRUST US OPPORTUNITIES A .00222 .00366 -.00442 .01944 .00431 -.00735 .01885 M&G NORTH AMERICAN .0034 .00371 -.0096 .02386* .00273 -.00754 .02593* M&G NORTH AMERICAN DIVIDEND .00407 .00435 -.00794 .02128 .00428 -.00706 .02298 MANSARTIS AMERIQUE -.00024 .00066 .00005 .0152 .00005 -.00108 .01354 MARLBOROUGH US MULTI-CAP .00453 .00473 -.0095 .01936 .00432 -.00753 .02021 MULTI MANAGER INVEST .00334 .00375 -.00961 .02573** .00355 -.00929 .02686** MULTIFONDO AMERICA A FI -.00215* -.00202* -.00954* .01288** -.00218* -.01076** .01181** MUTUAFONDO RENTA -.00015 -.00011 -.00113 -.00005 .0000 -.00167 .00004 NATIXIS ACTIONS US .00484** .00571** .00411 .01035 .00571** .00196 .00864 NINETY ONE AMERICAN FRANCHISE .0033 .00422 -.02205 .03653** .00343 -.02198 .0376** NN NORTH AMERICA .00481** .00516** .0015 .01474 .00474** .00213 .01547 NOBLE FUND -.00671** -.0059** -.01428 -.00442 -.00557** -.0162 -.00575 NORDEA INVEST NORTH AMERICA .00385 .0042 -.01789 .02863** .0035 -.01689 .02929** OHMAN ETISK INDEX USA A .00524** .00586** -.00237 .01934* .00547** -.00187 .01976* PALATINE AMERIQUE PALATINE .00379* .00381 -.00144 .00655 .00358 -.00044 .00788 PEKAO AMERICAN EQUITY .00002 .00058 -.00527 .00911 .0004 -.00479 .00957 PWM US DYNAMIC .00392 .00496 -.01395 .03564** .00492 -.01608 .03444** ROYAL LONDON US GROWTH .00304 .00349 -.00943 .02514* .00404 -.01226 .02596* SABADELL ESTADOS UNIDOS .00236 .00316 -.00101 .00921 .00212 .00115 .00943 SANTANDER ACCOES AMERICA A .00297 .0031 -.00438 .01221 .00257 -.0028 .01401 SANTANDER PF UNITED STATES .00473 .00496* -.01113 .0241* .00498* -.01099 .02521* SCHRODER QEP US CORE INST.INC .00525* .00535* -.01116 .02674** .00495* -.00993 .02793** SCHRODER US MID CAP A .00495* .00523* -.00107 .02043 .00513* -.00082 .02147 SCHRODER US SMALLER COS. A .0052* .00556* -.00512 .02459* .00537* -.00444 .02584* SCWID.AMER.GW.CL.A AC .00427 .00437 -.0082 .01809 .00418 -.00728 .0194 SEBINVEST NORDAMERIKA .00465* .00534* -.01145 .02361* .00551* -.01198 .0248* SKANDIA FONDER AB SF .00433* .00475* -.00929 .02163* .00462* -.00857 .0227* SMITH & WILLIAMSON NORTH AMERIC .00333 .00378 -.00651 .01789 .00392 -.00682 .01892 SPARINVEST INDEX USA .00469* .00564** -.01143 .0281** .00578** -.01292 .02722** SPP AKTIEFOND USA A .00539** .00581** -.0076 .02343* .00562** -.00677 .02436** ST JAMES'S PLACE NORTH AMERICAN .00586* .00631** -.00542 .02642* .00595* -.00522 .0281* STATE ST AUT NTH AMERICA .00499** .00534** .00027 .01387 .00507** .00108 .01474 STRAT INDICE USA(C) LEGAL .00399* .00428* .00092 .01194 .00412* .00176 .01274 SYDINVEST USA LIGEVAEGT .00598** .00582** -.00034 .01838 .00536* .00178 .0202 SYNCHRONY US EQUITY A .00302 .00348 -.0062 .01293 .00329 -.00602 .01329 TBF GLOBAL TECHNOLOGY .0024 .0049 -.00419 .01736 .00526 -.00572 .01488 THREADNEEDLE AM EXTENDED .00463 .00547* -.00552 .02008 .00509* -.00608 .02012 THREADNEEDLE AMERICAN RETAIL .00439 .00513* -.01328 .02756** .00491* -.01379 .02782** THREADNEEDLE AMERICAN SELECT RE .00339 .00421 -.00823 .02406* .00395 -.00955 .02412* THREADNEEDLE AMERICAN SMCOS .00385 .0044 -.0106 .02624* .00494 -.01163 .02706* 39 Appendix 1 –Alpha estimates for the first research hypothesis by model (continued) UBS (CH) EQUITY FUND .00291 .00352 -.0031 .02007* .00328 -.00443 .01946 UBS (CH) INST FD 2-EQ .0059*** .0063*** -.00065 .01589 .006** .00023 .01658 UNION INV.PRIVATFONDS .00308 .00382 -.00217 .01383 .00347 -.00122 .01407 US MULTI-FACTOR EQUITY .00512* .0052* -.01058 .02457* .00504* -.0094 .02578* US NEW TECHNOLOGY .00257 .00431 -.00232 .01363 .00387 -.00467 .01142 VANGUARD US EQUITY .0055** .00593** -.00079 .01413 .0057** .00016 .01489 XT USA .00407 .00419 -.0132 .02662* .00397 -.01215 .02797** ZIF AKTIEN USA A1 .00559* .00649* .00022 .03319** .00577* -.00169 .03209* ZIF AKTIEN USA PASSIV A1 .00682** .00754** .00715 .0289* .00702** .00568 .02838* ALLIANCE TRUST NORTH AMERICAN E .01651 .00972 -.03298 .01997 .01427 -.01656 .01042 ALLIANZ AMERIKA AANDELEN FONDS .00153 .00168 -.0038 .01626 .00171 -.00344 .01797 BARCLAYS US ALPHA A ACC DEAD - .00728* .00791* -.02886 .04779*** .00744* -.01544 .04584** CF GREENWICH AC. DEAD - Liquida .00978* .01017* -.02372 .04028 .01019* -.01498 .03936 CONNECT EQUITY USA GREEN I - TO .00769*** .0087*** -.00906 .02041 .00861*** -.00628 .01767 CONNECT EQUITY USA RED I DEAD - .00566** .00608** -.00924 .0166 .006** -.00409 .01613 CS (CH) US QUANT EQUITY FUND LI .0038 .00479 -.01802 .01194 .00442 .00398 .01148 DEUTSCHE (CH) I US EQUITIES LD .00643** .00699** -.01224 .03164** .00667** -.00503 .03085** DEXIA CLICKINVEST B INDEX LINKE .0016*** .00192*** .00923*** -.00348 .00177*** .00721** -.00229 DPAM INVEST B EQUITIES US DIVID .00434* .00439* -.0062 .01051 .00437* -.00429 .01164 EFG EQ.FUNDS NTH.AM. DEAD - Liq .00619* .00717* -.02042 .04257*** .00681* -.01296 .04262*** ETHOS EQUITIES NORTH AMERICA RP -.00599 -.00659 -.11703** .06444 -.00916 -.09341 .05363 EVLI FONDER ERIK PENSER AKTIEIN .00338 .00374 .05804* -.04243 .0036 .04211 -.03115 FF &.P US SMALL CAP. EQUITY B I .00854** .01065** -.02673 .0433** .01153** -.01672 .04546** HSBC PPUT NORTH AMERICAN DEAD - .00584 .0061 -.03183 .03326* .00562 -.01511 .03278* JYSKE INVEST US EQUITIES CL - T .00436 .00508* -.00834 .02626* .00448 -.00795 .02634* KAMES AMERICAN EQUITY B GBP - T .00213 .00288 -.02371 .02407 .00259 -.016 .0248 KBC EQUIMAX QUALITY STOCKS US 1 -.00078 -.00075 .01293 -.0102 -.00101 .01448 -.01076 LB.BL.INV.LINGOHR AMERIKA SYST -.00003 -.0009 -.02474 .02498 -.00033 -.02462 .02899* MAN GLG AMERICAN GROWTH RETAIL .00316 .00395 -.01369 .01789 .00492 -.01444 .01884 MITON AMERICAN RET ACC DEAD - L .00409 .0048 -.02945 .02795 .00496 -.01351 .02779 MULTI MANAGER HEALTH CARE - TOT .02289** .02489** .12522* .01032 .02004* .11176* .02211 MULTI MANAGER HEALTH CARE AKKUM .01149** .01088* .04983 .00747 .00899* .05055* .00846 MULTI MANAGER INVEST TEKN AKK - .00553 .00664 .01261 .01061 .00509 .01167 .01236 MULTI MANAGER INVEST TEKNOLOGI .01275 .0156* .06387 .01226 .0121 .05472 .02031 NEPTUNE US MAX ALPHA A DELISTED .0051 .0029 -.02222 .0539 .00116 .00277 .03876 OBJECTIF AMERIQUE $ COUV SOCIET -.00511*** -.00484*** -.00284 .00052 -.00537*** -.00171 .00013 OPPENHEIM KPL.AMER. EQUITIES - .00286 .00399 -.01352 .03016* .00496 -.0072 .02781 PIONEER INVESTMENTS AKTIEN DTL. -.00289 -.00161 -.00885 .00948 -.00233 -.01423 .00443 PUTM NORTH AMERICAN ACC DEAD - .00385 .00454 -.01547 .02542 .00487 -.01423 .02394 QUILVEST MULTI USA P QUILVEST G .00111 .00169 -.01043 .01013 .00195 -.00747 .01002 SCOTTISH WIDOWS HIFML US FOCUS .00378 .00393 -.01707 .01788 .00446 -.01398 .01879 SCOTTISH WIDOWS HIFML US STRATE .00435 .00474 -.01655 .02683 .006 -.01902 .02964 SIEMENS EQUITY NORTH AMERICA - .00296 .00313 -.00319 .02395* .00341 -.00446 .02576* SNS AM.AANDFDS. DEAD - Liquidat .00539** .00598** -.00044 .01821 .00552** .00033 .01897 SSGA US ALP.EQ.FD.I EUR STE.STR .00455 .00324 .00451 .0106 .00193 .01915 .00512 STAR LIBERTY AMERICA DEAD - Liq .00781** .00829** -.01593 .03183** .00797** -.00587 .03101** STOCKINDEX USA DEAD - Liquidate .00476* .00507* -.00822 .02718* .00473 -.0076 .02801** UBS US 130 30 EQUITY A DELISTED .0041 .00301 -.02757 .03911 -.00062 -.0193 .03544 UFF CROISSANCE PME X DEAD - Mer .00263 .00311 .00199 .02183 .00268 -.00968 .02324 UNIVERSAL INV.GESELL. HOTCHKIS& .00732 .00864 -.00617 .05638 .00762 -.01115 .05609 US SPECIAL EQUITY A DEAD - Liqu .00337 .00298 -.01307 .01863 .00222 -.00994 .02052 VOLKSBANK AM.INVEST T DEAD - Li .00367 .0036 -.00523 .02524 .00419 -.00572 .02778 VVA - AKTIEN USA F DEAD - Liqui .00543** .00598** -.00699 .01523 .00577** -.00326 .01561 WARBURG INVEST KPL. US DIVERSIF .0021 .00127 -.03376* .02266 .0015 -.02341 .0228 ABERDEEN NORTH AMERICAN EQUITY .00858** .00912** -.02316 .02768 .009** -.00202 .02824 AHORRO CORPORACION USA DEAD - M -.00139** -.00168** -.00354 -.00177 -.00176** -.00452 -.00093 ALLIANZ ACN.US COUVERT GLOBAL I -.00319* -.00384** .00116 -.01052 -.00372** .00012 -.00887 ALLIANZ ACTIONS US GLB. INVESTO .00512 .00499 -.00681 .01486 .00523 .00329 .01655 AMERICA LMM(C) LOUVRE (BANQUE D -.00193* -.00162 -.01237** .01035** -.00145 -.01173** .01001** AMERIQUE RENDEMENT (C) EDM.DE R .00211 .00246 -.00775 .00388 .00334 -.0101 .00685 AMUNDI AMERIKA BLUE CHIP STOCK -.00289* -.00328** -.01284* .01002 -.00346** -.01254* .01086 AMUNDI EQUITY STRATEGY USA A - .00272 .00313 -.00887 .0271* .00269 -.00787 .0275** AMUNDI FRN.INDEX USA I CAP - TO .00593* .00524 .00761 .00613 .00391 .01746 .00161 ANIMA GEO AMERICA A DEAD - Merg .0033 .00379 -.0033 .01451 .00337 -.00272 .01467 ANIMA NEW YORK MERGED SEE 88143 -.00335 -.0016 .02056 -.01085 -.00473* .01246 -.0081 APIUS AVENIR AMERIQUE DELUBAC A -.00759 -.00789 .01694 .00829 -.00833 .00503 .01448 AXA FRAMLINGTON EQ.INC. AC. - T .0065* .00736** .0057 .01897 .00738** -.0019 .02138 BANCAJA RN.VAR. ESTADOS UNIDOS .008* .00627 .00482 .02361 .00128 .01396 .015 BCY.ALPHASTARS US CAP. BARCLAYS .00319 .00312 .00447 .02279 .00112 .01537 .01729 BCY.AMERIQUE FCP WEALTH MANAGER .00314 .00312 .0035 .01225 .0017 .01183 .01049 BNP PARIBAS ACTS USA (C) BNP PA .00403 .00451 -.01852 .0252* .00407 -.01487 .02443* BNP PARIBAS B I EQUITY USA CAP .00482* .00576* -.00663 .01898 .00566* -.00522 .01784 BNP PARIBAS QUANTAMERICA BNP PA .00489 .0054 -.0036 .02694 .00489 .00223 .02676 CANDRIAM SUSTAINABLE NORTH AMER .00332 .00307 -.00515 .0095 .00328 -.0043 .01115 CEP GESTORA FONPENEDES BORSA US .00471* .00474* -.00707 .00982 .00481* -.00293 .00951 CNP ASSUR AMER(D) CAISSE NATION -.00075 -.00028 .00056 .00513 -.00053 -.00009 .00394 40 Appendix 1 –Alpha estimates for the first research hypothesis by model (continued) Note: This appendix contains the alpha estimates across different models for each one of the 285 mutual funds to answer the first research question. It supports the results presented in table 2 and it contains the winner and loser funds divided into two portfolios (the q-4 factor model column), to answer the second research question. *** Indicates statistical significance at the 1% level. **Indicates statistical significance ate the 5% level. *Indicates statistical significance at the 10% level. COLUMBIA SECURITIES DEAD - Merg -.00169 .00012 -.01178 .03476 -.00036 -.00766 .03093 CPR ACTIVE US (H - EUR) - P - T -.0031*** -.00319*** -.00884** .00177 -.00327*** -.00727* .00111 CSIF NORTH AMERICA INDEX BLUE - .01015*** .01056*** -.01668 .04089*** .01001*** -.00617 .04006*** DWS INVESTMENT US AKTIEN TYP O .00251 .00266 -.01877 .02651** .00186 -.01648 .02824** ECHIQUIER AMERIQUE FINANCIERE D .00824** .00883** -.01103 .02926 .00876** -.00843 .03286* EPARAMERIC(C) LA POSTE - TOT RE .00362 .00388 -.01133 .02467 .00285 .00125 .02337 EQUITY STRATEGY NA.T DEAD - Mer .00704* .00736* -.0241 .03964** .00706* -.01643 .04156*** EURIZON AZIONI PMI AMERICA - TO .00207 .00278 -.02021* .02026 .00282 -.01821 .022* EUROVALOR ESTADOS UNIDOS FI - T .00243 .00305 -.00521 .01519 .00284 -.00496 .0159 FD.MAN.SWITZ.AG CH INST 2 EQUIT .00672** .00797** -.00442 .02618* .00791** -.00474 .02532* FONCAIXA CARTERA BOLSA USA FI ( .00522 .00477 .00968 .01363 .004 .02153 .01285 FONCAIXA I BOLSA USA FI DEAD - -.002* -.00211** -.00588 .00489 -.00285*** -.00289 .00329 FONCAIXA USA FI (IN MER) DEAD - -.00407 -.00369 .00591 .00625 -.00495* .01254 .0057 GROUPAMA ACTIONS MID CAP MERGED .00781 .00768 .00573 .01997 .00484 .01555 .01569 GROUPAMA US STOCK EURO BANQUE F .00819* .00801 .00911 .01874 .00422 .01613 .01472 GROUPAMA USACTIONS (C) MERGED S .00714 .0072 .01225 .01234 .0054 .02043 .0097 HANSAINVEST HANSEATISCHE INV.GE .00439 .00366 -.00433 .03667 .00159 .00594 .03249 HSBC FSAVC NORTH AMERICA DEAD - .00516 .00533 -.02971 .02636 .00527 -.01615 .02785 INDOSUEZ ELITE US CREDIT AGRICO .00539* .00585* -.0114 .02835* .00479 -.00246 .02621* INVERSEGUROS SEGURFONDO USA - T .00111 .00111 .00427 -.00233 .00049 .00375 -.00273 IQAM EQUITY US (RT) DEAD - Merg .00312 .00353 -.00701 .01943 .0028 -.00547 .01999 JPM US A ACC DEAD - Merged .00374 .00342 -.0225 .02307 .00378 -.01741 .02413 JYSKE INVEST USA AKTIER KL - TO .00369 .00433* -.00817 .02472** .00404 -.00797 .02546** KBC INDEX FD. USA CAP DEAD - Me .00549** .00589** -.01135 .01706 .00574** -.00639 .01638 KUTXAVALOR EEUU FI DEAD - Merge .00607 .0046 .01639 .00946 .00351 .03094 .002 MAITRE AMERICAN EQUITIES ASST.A .00198 .00304 -.01425 .02538* .00348 -.01784 .02506* MAM AMERIQUANT DEAD - Merged .00007 .00063 -.01277 .00176 .0008 -.00944 .00241 MARTIN CURRIE NTH.AMER. CL.A - .00714* .00774** -.01995 .03537** .00702* -.00939 .03618** MARTIN MAUREL COMPOSITION AMERI .00334 .00392 -.0015 .01787 .00372 -.00163 .01839 MC FDF AMERICA A MERGED SEE 257 -.00686*** -.00556** -.01348 .00639 -.00754*** -.01084 .0041 MULTI MANAGER INVEST USA .00327 .00356 -.01036 .02636** .00351 -.00971 .02772** NB ACOES AMERICA DEAD -.00267 -.00158 -.01198 .0053 -.00182 -.01013 .00429 NEUFLIZE USA OPNS.$ C OBC ASST. -.00176 .00025 -.00584 .02712 .00064 -.01083 .02842 OBJECTIF GEST VAL AMER. SOC DE .00181 .00273 -.01108 .01836 .00232 -.00286 .01742 ODDO BHF ACTIONS USA CR-EUR - T -.00235*** -.00245*** -.00212 -.0005 -.00258*** -.00198 -.00067 OPTIMIX AMERICA FD DEAD - Merge .00548* .00479 -.01517 .02236 .00524* -.01401 .02616* PKO AKCJI RYNKU AMERYKANSKIEGO -.00159 -.00146 -.00874 .00556 -.0017 -.00222 .00547 R-CO CONVICTION USA C DEAD - Me .00063 .00109 -.00434 .01243 .00077 -.00399 .01335 RENTA 4 TECNOLOGIA DEAD - Merge .00212 .00218 -.00356 .00609 .00213 -.00089 .0078 SCWID.UK.SMCOS.CL.B AC. DEAD - .00449 .00563* -.00528 .01048 .00558 -.00721 .01041 SLGP PRIGEST US DEAD - Merged .00319 .00375 .00373 .00823 .00578 -.00233 .01144 SLI IGNIS AMERICAN GROWTH INC - .00541 .00724* -.02657 .02957 .00717* -.00797 .02716 SOCIETE GENERALE ACTIONS US SEL .00171 .00233 -.00674 .01123 .00213 -.0058 .01258 SPARINVEST VALUE USA KL DEAD - .00291 .00309 -.0038 .01259 .00268 -.00242 .01428 SSGA NA.ENH.EQ.FD.I STE.STR.GLB .00874*** .00907*** -.01437 .03187** .00848*** -.00213 .03146** SSGA US IDX.EQ.FD.P USD STE.STR .00664** .00734** -.01566 .02125 .00689** -.0019 .02122 SWC (CH) INDEX EF MSCI USA A - .00588 .00409 .01447 .01026 .00245 .022 .00689 SYNERGIA AZIONARIO USA DEAD - M .00217 .00307 .00825 .00643 .00079 .00407 .01024 Santander Accoes USA -.00789* -.00644 .00366 -.01242 -.00761* .00029 -.01382 THE WESTCHESTER US DOLLAR ACC - .00693* .0079* -.00873 .02445 .00776* -.00918 .02126 TOBAM ANTI BENCHMARK US EQUITY .00946** .00946** .00000 .04017** .00976*** .00195 .04009** UFF CROISSANCE PME X DEAD - Mer .00263 .00311 .00199 .02183 .00268 -.00968 .02324 WARBURG INVEST KPL. WARBURG AME .00431 .00271 -.00012 .01448 .00223 .01776 .00815 ALPHA COSMOS STARS USA EQUITIES .0026 .00337 -.00544 .01744 .00325 -.00596 .01807 ALTA USA DEAD - Merged 9104PL -.00058 -1.00e-05 -.01018 .01719 -.00031 -.01092 .01633 Aktia America B .00266 .00287 -.00147 .01598 .00276 -.00115 .01705 BCV SYSTEMATIC PREMIA US EQUITY .00503** .00553** .00032 .01036 .00539** .00186 .01169 FIM USA - TOT RETURN IND .00003 .00096 -.00916 .01666 -.0002 -.00844 .01592 FoLocalTapiola ESG USA Mid lha7 .00241 .00297 .00247 .00997 .00219 .00377 .00993 FoleQ USA Indeksi 1 Kha4 .00504** .00544** -.00175 .01572 .00501** -.00064 .01635 NN SUB SPOLEK DYWIDENDOWYCH USA -.00419** -.00495*** .0133 -.01427* -.00509*** .01332 -.01422 NORDEA MEDICA KASVU DEAD - Merg .00916** .00982** .03242 .01952 .00891** .03273 .019 Nordea North American Dividend .00386 .00412 -.01835 .02718** .00334 -.01721 .0277** Piraeus US Equity Fund R .00235 .00229 .00298 .00718 .00222 .00415 .00831 SEB North America Index B .00446** .00481** .00077 .01406 .00457** .0015 .01493 TRITON AMERICAN EQUITY INTERNAT -.00005 .0015 -.00413 .01207 .00066 -.00466 .01232 41 Appendix 2 –Alpha estimates for the robustness check by model Mutual fund name Carhart 4factor model q-4 factor model Carhart 4-factor model plus BW q-4 factor model plus BW 3 BANKEN AMERIKA .00302 .00382 .0033 .00428 ACOMEA AMERICA A1 .00016 .0003 .00016 .00074 ALLIANZ AZIONI AMERICA .00088 .00123 .00089 .00133 ALLIANZ RCM US EQUITY A .00184 .00223 .00165 .00221 AMERI GAN D GROUPAMA .00333 .00377 .0037 .00424 AMONIS EQUITY US CANADA .00505** .0062** .00518* .00617** AMUNDI AKTIEN ROHSTOFFE A -.00599 -.00563 -.00742* -.00626 AMUNDI AZIONARIO .00055 .00085 .00126 .00157 ANIMA AMERICA A .00371 .00405 .00389 .00428 ARNIKA ARVEST AM.STARS .00352 .00361 .00316 .00371 ASI AMER.UCND.EQ. RET .0021 .0029 .00204 .00292 ASI AMERICAN EQUITY .00206 .0019 .00199 .00199 ASI STANDARD LIFE NORTH AMERICA .00288 .00348 .00288 .00361 AVIVA AMERIQUE I .00266 .00323 .00235 .00293 AXA FRAMLINGTON GLOBAL TECHNOLO .00383 .00622 .00296 .00516 AXA INDICE USA A C .00441* .00456* .00426 .0045* AXA ROSENBERG AMERICAN .00353 .00377 .00362 .00404 BANKIA BOLSA USA .00019 .00047 .00004 .00038 BANKIA INDICE S&P 500 -.00187*** -.00214*** -.00184*** -.00212*** BATI ENTREPRENDRE .0023 .00296 .00211 .00289 BBVA BOLSA USA .002 .00241 .00192 .00242 BBVA INDICE USA PLUS -.00398*** -.00393*** -.00383*** -.00378*** BCV ENHANCED US -.00169*** -.00178*** -.00163*** -.00172*** BLACKROCK US DYNAMIC INC .00306 .00298 .00338 .00341 BLACKROCK US OPPS.AC .00218 .00314 .00171 .00291 BLLE.GIFF.AMER.A AC. .00536 .00746** .00511 .00706* BMO NORTH AMERICAN EQUITY 1 .00371 .00361 .00356 .00366 BMO US SMALLER COMPANIES .00439 .0048 .00488 .00565 BNL AZIONI AMERICA .00178 .0026 .00202 .00273 BNY MELLON US OPPORTUNITIES .00104 .00252 .00102 .00253 BSO AMER(C) SEGIPA .00073 .00152 .00072 .00158 C-QUADRAT GLOBAL EQUITY ESG R - .00134 .0009 .00056 -.00002 CAIXA ACOES EUA .00306 .00343 .00268 .00314 CAIXABANK BOL SELECCION USA .00168 .00215 .00162 .00219 CAIXABANK BOLSA USA .00263 .00271 .00266 .00289 CAJA INGENIEROS BOLSA .00178 .00168 .00136 .00128 CD AMERIQUE STGIS.A .00207 .00279 .00195 .00279 CLEMED.INV.IOM US EQ .00407 .00415 .00425 .00454 CM-AM INDICIEL AMERIQUE 500 C .00382 .0041 .00374 .00411 COLUMBUS AMERICAN -.00184 -.00185 -.00104 -.0009 COLUMBUS US MARKET -.00458* -.00424 -.00456 -.00403 COVEA ACTIONS AMERIQUE A .00266 .00323 .00235 .00293 CPR USA ESG - P (C) .00369 .00379 .00378 .004 DANSKE INVEST BIOTEKNOLOGI .00754 .01247** .00604 .01103* DANSKE INVEST TEKNOLOGI .00452 .0071** .00314 .00527 DANSKE INVEST USA -.00213** -.00237** -.00189* -.00198* DANSKE INVEST USA DKK D .0039 .00396 .00405 .00436 DIT BIOTECHNOLOGIE .00397 .00964* .00598 .01235* DPAM CAPITAL .0046* .00488* .00455* .00494* DPAM CAPITAL B .00561** .00546** .00593** .00595** DPAM CAPITAL B EQUITIES US .0029 .00287 .00311 .00324 DPAM DBI-RDT B .00438 .00485 .00414 .00488 DWS INV.NORDAMERIKA .00327 .00467* .00369 .00518* EDMOND DE ROTHSCHILD .00336 .00445* .00326 .00425 EOLE (C) .0033 .00415 .00269 .00352 ERSTE BEST OF AMERICA .00242 .00314 .00195 .00289 ESPA STOCK AMERICA .00155 .00189 .00042 .00096 ESSOR USA OPPORTUNITIES P .0034 .00482 .00299 .00447 ETOILE ACTIONS US .00422* .00451* .00418 .00452 ETOILE MULTI GESTION .00188 .00249 .00177 .00239 Panel B: Robustness check mutual fund's alpha 42 Appendix 2 –Alpha estimates for the robustness check by model (continued) ETOILE MULTI GESTION ETATS-UNIS -.00282*** -.00285*** -.00265*** -.00266*** EURIZON AM AZIONI .00062 .00078 .00154 .00179 EURIZON AZIONI AMERICA .00201 .00239 .00236 .00286 FEDERAL INDICIEL US .0043* .00456* .00437* .00478* FID.AM.INSTL.FD. .00388 .00392 .00396 .00419 FIDELITY AMERICAN A ACC .00227 .00306 .00195 .00294 FIDELITY AMERICAN SPECIAL .00455 .00442 .00475 .00495 FIDEURAM MASTER SEL.EQ .00165 .00206 .00193 .00243 FONDMAPFRE BOLSA AMERICA .00268 .00215 .00248 .00202 FOURPOINTS AMERICA .00168 .00188 .00125 .00158 GAM NORTH AMERICAN GROWTH A .00561* .00594* .00552 .00605 GROUPAMA US EQUITIES .00361 .00388 .00365 .00391 GUT.KAPITALANLAGE .00152 .00218 .00114 .00206 GUTZWILLER FON.MAN.ONE .00198 .00244 .00161 .00228 HALIFAX NORTH AMERICAN C .00423 .00429 .00433 .00458 HANDELSINVEST NORDAMERIKA .00406 .00449 .00469 .00498 HSBC AMERICAN INDEX .00462 .00482 .00451 .00489 IBERCAJA BOLSA USA .00251 .00235 .00251 .00247 ING DIRECT FONDO NARANJA .00397 .00413 .00417 .00433 INVESCO US EQUITY (UK) .00223 .00254 .00255 .00326 INVESTITORI AMERICA .00323 .00324 .00344 .00362 IQAM QUALITY EQUITY .00424 .00383 .00429 .00397 JANUS HENDERSON INST .00354 .00388 .0032 .00374 JANUS HENDERSON US GROWTH A .00172 .00289 .00169 .00291 JPM US EQUITY .00557* .00533* .00553 .00559 JPM US SELECT .00335 .00362 .00357 .00396 JPM US SMALL CAP GROWTH A .00318 .00597 .00297 .00577 JUPITER NTH AMERICAN INC .00372 .00358 .00352 .00368 KATHREIN US-EQUITY .00281 .00388 .00295 .00428 KBC EQ.FD.BUYBACK AM .0041* .00423* .00411 .00443 KBC EQ.FD.SMALL CAPS .00417 .00418 .00424 .00435 KBC EQ.FD.US VALUE CAP .00446* .00437* .00443* .00451* KBC EQUITY NORTH AMERICA .00344 .00392 .00338 .00394 KBC INSTL.FD.US EQ. CAP .0046* .00503* .00447 .00495* KEPLER US AKTIENFONDS A .0046 .00394 .00557* .00502 KLP AKSJE USA INDEKS .00462* .00492* .00454* .00498* KLP AKSJE USA INDEKS II -.00343* -.00274 -.00421** -.00318* KUTXABANK BOLSA EEUU -.00197** -.0026*** -.00197** -.0026*** LABORAL KUTXA BOLSA -.00379*** -.00389*** -.00342*** -.0036*** LANSFORSAKRINGAR .00483* .00501* .00487* .00523* LANSFORSAKRINGAR USA .00409 .00451 .00482 .00538* LAZARD ACTIONS .00204 .00217 .00194 .0022 LBPAM ISR ACTIONS AMERIQUE .00172 .0023 .00135 .00208 LIONTRUST US OPPORTUNITIES A .00139 .00353 .00172 .00392 M&G NORTH AMERICAN .00333 .00293 .00303 .00279 M&G NORTH AMERICAN DIVIDEND .00351 .00377 .00273 .00321 MANSARTIS AMERIQUE -.00112 -.00104 -.0013 -.00141 MARLBOROUGH US MULTI-CAP .00348 .00325 .00399 .00389 MULTI MANAGER INVEST .00275 .00305 .00251 .00297 MULTIFONDO AMERICA A FI -.00211* -.00225* -.00193 -.00219* MUTUAFONDO RENTA .00011 .00026 -.00013 1.00e-05 NATIXIS ACTIONS US .00468* .00569** .00478* .00566* NINETY ONE AMERICAN FRANCHISE .00276 .00253 .00227 .00207 NN NORTH AMERICA .0043* .00448* .00417 .00443* NOBLE FUND -.00656** -.0054* -.01074*** -.00991*** NORDEA INVEST NORTH AMERICA .00334 .00303 .00379 .00355 OHMAN ETISK INDEX USA A .0048* .00517** .0049* .00536* PALATINE AMERIQUE PALATINE .00366 .0036 .00329 .0034 PEKAO AMERICAN EQUITY -.00043 -1.00e-05 -.00067 -.00016 PWM US DYNAMIC .00336 .00426 .00346 .00433 ROYAL LONDON US GROWTH .0023 .00341 .00294 .00428 SABADELL ESTADOS UNIDOS .00194 .00181 .00217 .00203 SANTANDER ACCOES AMERICA A .00279 .00263 .00261 .00261 SANTANDER PF UNITED STATES .00447 .00487 .00478 .00542 SCHRODER QEP US CORE INST.INC .0049* .00478 .00476 .00482 SCHRODER US MID CAP A .00428 .00453 .0043 .00467 SCHRODER US SMALLER COS. A .00428 .00463 .00423 .00472 SCWID.AMER.GW.CL.A AC .00394 .00403 .0041 .0044 SEBINVEST NORDAMERIKA .00417 .00516* .00419 .0054* SKANDIA FONDER AB SF .00411 .0045 .00445 .00507* SMITH & WILLIAMSON NORTH AMERIC .00264 .00327 .00246 .00328 43 Appendix 2 –Alpha estimates for the robustness check by model (continued) SPARINVEST INDEX USA .00406 .00514* .00442 .0055* SPP AKTIEFOND USA A .00501* .00534** .00508* .00558* ST JAMES'S PLACE NORTH AMERICAN .00515 .00548* .0053 .00588 STATE ST AUT NTH AMERICA .00449* .00471* .00425* .00457* STRAT INDICE USA(C) LEGAL .00356 .00389 .0032 .00362 SYDINVEST USA LIGEVAEGT .00563* .00532* .00527* .00514 SYNCHRONY US EQUITY A .00252 .00292 .00248 .00294 TBF GLOBAL TECHNOLOGY .00347 .00572 .00189 .00385 THREADNEEDLE AM EXTENDED .00411 .00487 .00492 .00579* THREADNEEDLE AMERICAN RETAIL .00388 .00452 .00414 .00489 THREADNEEDLE AMERICAN SELECT RE .00278 .00354 .00321 .00406 THREADNEEDLE AMERICAN SMCOS .0033 .00431 .00258 .00368 UBS (CH) EQUITY FUND .0022 .00289 .00241 .00308 UBS (CH) INST FD 2-EQ .00547** .00571** .00535** .00568** UNION INV.PRIVATFONDS .0025 .0029 .00272 .00321 US MULTI-FACTOR EQUITY .0046 .00464 .00466 .00491 US NEW TECHNOLOGY .00187 .00299 .00175 .00263 VANGUARD US EQUITY .00506** .00539** .00487* .00531** XT USA .00388 .00394 .00445 .00477 ZIF AKTIEN USA A1 .00548 .0059 .00467 .00475 ZIF AKTIEN USA PASSIV A1 .00635* .00676* .00541 .00554 ALLIANCE TRUST NORTH AMERICAN E .01651 .01427 .01004 .00698 ALLIANZ AMERIKA AANDELEN FONDS .00106 .0013 .00073 .00112 BARCLAYS US ALPHA A ACC DEAD - .00728* .00744* .0073* .0077* CF GREENWICH AC. DEAD - Liquida .00978* .01019* .00849 .0092 CONNECT EQUITY USA GREEN I - TO .00769*** .00861*** .00799** .0088*** CONNECT EQUITY USA RED I DEAD - .00566** .006** .00533* .0058** CS (CH) US QUANT EQUITY FUND LI .0038 .00442 .00348 .00432 DEUTSCHE (CH) I US EQUITIES LD .00643** .00667** .00634* .00681** DEXIA CLICKINVEST B INDEX LINKE .0016*** .00177*** .00122* .00143** DPAM INVEST B EQUITIES US DIVID .00434* .00437* .00449 .0046 EFG EQ.FUNDS NTH.AM. DEAD - Liq .00619* .00681* .00574 .00642* ETHOS EQUITIES NORTH AMERICA RP -.00599 -.00916 -.00555 -.00944 EVLI FONDER ERIK PENSER AKTIEIN .00338 .0036 -.00221 -.00195 FF &.P US SMALL CAP. EQUITY B I .00854** .01153** .00834** .01167** HSBC PPUT NORTH AMERICAN DEAD - .00584 .00562 .00546 .00549 JYSKE INVEST US EQUITIES CL - T .00396 .00412 .00421 .0044 KAMES AMERICAN EQUITY B GBP - T .00213 .00259 .00224 .00316 KBC EQUIMAX QUALITY STOCKS US 1 -.00078 -.00101 -.00073 -.00084 LB.BL.INV.LINGOHR AMERIKA SYST -.00003 -.00033 -.00054 -.00066 MAN GLG AMERICAN GROWTH RETAIL .00316 .00492 .00284 .00478 MITON AMERICAN RET ACC DEAD - L .00409 .00496 .00327 .00438 MULTI MANAGER HEALTH CARE - TOT .02289** .02004* .02828** .02641** MULTI MANAGER HEALTH CARE AKKUM .01149** .00899* .01258** .01086* MULTI MANAGER INVEST TEKN AKK - .00553 .00509 .00461 .00423 MULTI MANAGER INVEST TEKNOLOGI .01275 .0121 .01486* .01444 NEPTUNE US MAX ALPHA A DELISTED .0051 .00116 .00401 -.00024 OBJECTIF AMERIQUE $ COUV SOCIET -.00511*** -.00537*** -.00476*** -.00504*** OPPENHEIM KPL.AMER. EQUITIES - .00286 .00496 .00259 .0048 PIONEER INVESTMENTS AKTIEN DTL. -.00289 -.00233 -.00366 -.00363 PUTM NORTH AMERICAN ACC DEAD - .00385 .00487 .00401 .00519 QUILVEST MULTI USA P QUILVEST G .00111 .00195 .00065 .00165 SCOTTISH WIDOWS HIFML US FOCUS .00378 .00446 .0032 .00413 SCOTTISH WIDOWS HIFML US STRATE .00435 .006 .0048 .00695 SIEMENS EQUITY NORTH AMERICA - .0025 .00317 .00224 .00318 SNS AM.AANDFDS. DEAD - Liquidat .00493* .00509* .00524* .00549* SSGA US ALP.EQ.FD.I EUR STE.STR .00455 .00193 .00416 .00201 STAR LIBERTY AMERICA DEAD - Liq .00781** .00797** .00765** .00791** STOCKINDEX USA DEAD - Liquidate .00415 .00424 .0042 .00446 UBS US 130 30 EQUITY A DELISTED .0041 -.00062 .00305 -.00143 UFF CROISSANCE PME X DEAD - Mer .00263 .00268 .00154 .00149 UNIVERSAL INV.GESELL. HOTCHKIS& .00732 .00762 .00951 .00961 US SPECIAL EQUITY A DEAD - Liqu .00374 .00275 .00404 .00317 VOLKSBANK AM.INVEST T DEAD - Li .00367 .00419 .00326 .00417 VVA - AKTIEN USA F DEAD - Liqui .00543** .00577** .00521* .00563** WARBURG INVEST KPL. US DIVERSIF .0021 .0015 .00003 -.00032 ABERDEEN NORTH AMERICAN EQUITY .00858** .009** .00822** .00907** AHORRO CORPORACION USA DEAD - M -.00139** -.00176** -.00131* -.00161** ALLIANZ ACN.US COUVERT GLOBAL I -.00319* -.00372** -.0033* -.00363** ALLIANZ ACTIONS US GLB. INVESTO .00512 .00523 .00435 .00482 AMERICA LMM(C) LOUVRE (BANQUE D -.00193* -.00145 -.00183* -.00133 44 Appendix 2 –Alpha estimates for the robustness check by model (continued) AMERIQUE RENDEMENT (C) EDM.DE R .00211 .00334 .00169 .00313 AMUNDI AMERIKA BLUE CHIP STOCK -.00273* -.00314** -.00221 -.0025 AMUNDI EQUITY STRATEGY USA A - .00192 .002 .00156 .0017 AMUNDI FRN.INDEX USA I CAP - TO .00593* .00391 .00514 .00343 ANIMA GEO AMERICA A DEAD - Merg .00305 .00333 .00311 .00343 ANIMA NEW YORK MERGED SEE 88143 -.00335 -.00473* -.00473* -.00618** APIUS AVENIR AMERIQUE DELUBAC A -.00759 -.00833 -.01208** -.01303** AXA FRAMLINGTON EQ.INC. AC. - T .0065* .00738** .00572 .00677* BANCAJA RN.VAR. ESTADOS UNIDOS .008* .00128 .00967** .00257 BCY.ALPHASTARS US CAP. BARCLAYS .00319 .00112 .00217 .00031 BCY.AMERIQUE FCP WEALTH MANAGER .00314 .0017 .00209 .00089 BNP PARIBAS ACTS USA (C) BNP PA .00403 .00407 .00396 .00388 BNP PARIBAS B I EQUITY USA CAP .00482* .00566* .00442 .00535* BNP PARIBAS QUANTAMERICA BNP PA .00489 .00489 .00408 .00397 CANDRIAM SUSTAINABLE NORTH AMER .00286 .00288 .00318 .00347 CEP GESTORA FONPENEDES BORSA US .00471* .00481* .00475* .00491* CNP ASSUR AMER(D) CAISSE NATION -.00071 -.0005 -.00124 -.00124 COLUMBIA SECURITIES DEAD - Merg -.00169 -.00036 -.00355 -.00238 CPR ACTIVE US (H - EUR) - P - T -.0031*** -.00327*** -.0028*** -.00291*** CSIF NORTH AMERICA INDEX BLUE - .01015*** .01001*** .00955*** .00965*** DWS INVESTMENT US AKTIEN TYP O .00213 .00146 .00218 .00161 ECHIQUIER AMERIQUE FINANCIERE D .00824** .00876** .00767* .00828* EPARAMERIC(C) LA POSTE - TOT RE .00362 .00285 .00222 .00194 EQUITY STRATEGY NA.T DEAD - Mer .00704* .00706* .00665* .0069* EURIZON AZIONI PMI AMERICA - TO .00207 .00282 .00179 .00257 EUROVALOR ESTADOS UNIDOS FI - T .00206 .00253 .00202 .0026 FD.MAN.SWITZ.AG CH INST 2 EQUIT .00672** .00791** .00666** .00789** FONCAIXA CARTERA BOLSA USA FI ( .00522 .004 .00445 .00366 FONCAIXA I BOLSA USA FI DEAD - -.002* -.00285*** -.00205* -.00302*** FONCAIXA USA FI (IN MER) DEAD - -.00407 -.00495* -.00479 -.00543* GROUPAMA ACTIONS MID CAP MERGED .00781 .00484 .00602 .00295 GROUPAMA US STOCK EURO BANQUE F .00819* .00422 .0075 .00385 GROUPAMA USACTIONS (C) MERGED S .00714 .0054 .00508 .00318 HANSAINVEST HANSEATISCHE INV.GE .00439 .00159 .00262 .00005 HSBC FSAVC NORTH AMERICA DEAD - .00516 .00527 .0049 .00532 INDOSUEZ ELITE US CREDIT AGRICO .00539* .00479 .00476 .00423 INVERSEGUROS SEGURFONDO USA - T .00111 .00049 .00023 -.0005 IQAM EQUITY US (RT) DEAD - Merg .00276 .0028 .00176 .00185 JPM US A ACC DEAD - Merged .00374 .00378 .00394 .0042 JYSKE INVEST USA AKTIER KL - TO .00319 .00361 .00355 .0041 KBC INDEX FD. USA CAP DEAD - Me .00549** .00574** .00538* .00574* KUTXAVALOR EEUU FI DEAD - Merge .00607 .00351 .00507 .00273 MAITRE AMERICAN EQUITIES ASST.A .00198 .00348 .0017 .00314 MAM AMERIQUANT DEAD - Merged .00007 .0008 -.00122 -.00048 MARTIN CURRIE NTH.AMER. CL.A - .00714* .00702* .00675* .00675* MARTIN MAUREL COMPOSITION AMERI .00306 .00353 .0029 .00352 MC FDF AMERICA A MERGED SEE 257 -.00686*** -.00754*** -.00753*** -.00831*** MULTI MANAGER INVEST USA .00271 .00298 .00246 .00289 NB ACOES AMERICA DEAD -.00267 -.00182 -.00262 -.00163 NEUFLIZE USA OPNS.$ C OBC ASST. -.00176 .00064 -.00379 -.00153 OBJECTIF GEST VAL AMER. SOC DE .00181 .00232 .00148 .00229 ODDO BHF ACTIONS USA CR-EUR - T -.00235*** -.00258*** -.00201*** -.00223*** OPTIMIX AMERICA FD DEAD - Merge .00507 .00504 .00534 .00588* PKO AKCJI RYNKU AMERYKANSKIEGO -.00159 -.0017 -.00236 -.00225 R-CO CONVICTION USA C DEAD - Me .00002 .00019 -.00017 .00012 RENTA 4 TECNOLOGIA DEAD - Merge .00254 .0025 .00257 .00276 45 Appendix 2 –Alpha estimates for the robustness check by model (continued) Note: This appendix contains the alpha estimates across different models for each one of the 285 mutual funds, regarding the robustness check provided for the first research question. It supports the results presented in table 3. *** Indicates statistical significance at the 1% level. **Indicates statistical significance ate the 5% level. *Indicates statistical significance at the 10% level. SCWID.UK.SMCOS.CL.B AC. DEAD - .00341 .00449 .00342 .00456 SLGP PRIGEST US DEAD - Merged .00319 .00578 .00246 .00517 SLI IGNIS AMERICAN GROWTH INC - .00541 .00717* .00533 .00727 SOCIETE GENERALE ACTIONS US SEL .00077 .00104 .00032 .00067 SPARINVEST VALUE USA KL DEAD - .00291 .00268 .00315 .00289 SSGA NA.ENH.EQ.FD.I STE.STR.GLB .00874*** .00848*** .00841*** .00829** SSGA US IDX.EQ.FD.P USD STE.STR .00664** .00689** .00618* .00658** SWC (CH) INDEX EF MSCI USA A - .00588 .00245 .00415 .00076 SYNERGIA AZIONARIO USA DEAD - M .00217 .00079 .00105 -.00019 Santander Accoes USA -.00755 -.00726 -.00685 -.00665 THE WESTCHESTER US DOLLAR ACC - .00693* .00776* .00643 .00729 TOBAM ANTI BENCHMARK US EQUITY .00946** .00976*** .00975** .01032*** UFF CROISSANCE PME X DEAD - Mer .00263 .00268 .00154 .00149 WARBURG INVEST KPL. WARBURG AME .00431 .00223 .00336 .00172 ALPHA COSMOS STARS USA EQUITIES .00223 .00293 .00205 .00285 ALTA USA DEAD - Merged 9104PL -.00118 -.00081 -.00096 -.00069 Aktia America B .00197 .00222 .00147 .00189 BCV SYSTEMATIC PREMIA US EQUITY .00462* .00509* .00437 .00498* FIM USA - TOT RETURN IND -.00043 -.0006 -.00052 -.00082 FoLocalTapiola ESG USA Mid lha7 .00154 .00163 .00169 .00181 FoleQ USA Indeksi 1 Kha4 .00462* .00473* .00453* .00471* NN SUB SPOLEK DYWIDENDOWYCH USA -.0041** -.00501** -.00543*** -.00649*** NORDEA MEDICA KASVU DEAD - Merg .00916** .00891** .00931* .00936* Nordea North American Dividend .00347 .00302 .00388 .00349 Piraeus US Equity Fund R .00184 .00188 .00179 .002 SEB North America Index B .00392* .00415* .00385 .00418 TRITON AMERICAN EQUITY INTERNAT -.0002 .00075 -.00073 .00024