Active management in mutual funds with concentrated benchmarks: A major dilemma to fulfill the EU directive on portfolio concentration limits.

his doctoral thesis is focused on the increasing importance of mutual funds as saving vehicles for the finances of individuals in the economies of developed countries over the last half of the twentieth century. Among them, the equity funds that take the form of open-end funds like mutual funds are becoming one of the most relevant instruments. Mutual funds collect money from investors and invest it in stocks, bonds, other funds, and so on, and their performance depends on the mix of the allocated securities. Its main objectives were to contribute to the topic of the Active Share (AS) as applied to mutual funds in Eurozone countries. This investigation was mainly motivated by the need to analyse how both the level of concentration of the benchmarks and the level of concentration of the mutual fund industries in the Eurozone affect active management. Due to the scarce number of studies focused on this topic in Europe our results have relevant implications to investors, market supervisors, and policymakers. In Chapter 1, we analysed how article 52 in Directive 2009/65/EC (UCITS IV) on risk of portfolio diversification could distort the accuracy of AS due to the higher level of concentration in Eurozone domestic benchmarks. In this chapter, we were able to identify truly active management in Eurozone mutual funds. Thus, we developed an algorithm to capture the spurious AS (sAS), which is defined as the minimum AS, that is not a consequence of active decisions made by equity fund managers. The results provide evidence that the Directive negatively influences the accuracy of the AS shown for managers who work with very concentrated domestic Eurozone benchmarks such as the PSI 20 (Portugal), ATX 20 (Austria), and IBEX 35 (Spain). In contrast, the evidence from the sAS for the least concentrated domestic Eurozone benchmarks shows that AS reported by managers who work in France, Germany, and Finland are much more accurate. Hence, these results prove that direct AS comparisons in the Eurozone are not feasible and lead us to obtain three AS thresholds per domestic equity benchmark, which are the minimum values of AS needed to confirm that domestic equity funds are significantly active at 90%, 95%, and 99% confidence levels.We also analysed the level of active management in the domestic equity funds registered in each Eurozone country. Our findings also show that the high concentration level and the heterogeneity present in the domestic equity funds in the Eurozone prevent the direct comparability of the AS. Therefore, we had to consider the level of AS over the spurious level and the characteristics of each market that produces significant and different styles of active management. For that, we formulated an actual active share (aAS) that considered the level of concentration in the domestic equity funds of the Eurozone markets and the limits of the portfolio concentration on European regulation. We define aAS as the difference between the monthly AS obtained for each domestic equity fund minus its monthly AS threshold obtained previously at 90%, 95%, and 99% confidence levels. Focusing on the most relevant mutual fund industries in the Eurozone countries, we find that France is the most active domestic equity fund market. Spanish and Italian markets also show high levels of actual active management despite the large concentration in their domestic benchmarks. Conversely, domestic equity funds registered in Germany show lower levels of active management. Our research results support the hypothesis that aAS corrects the potential bias in the original AS caused by both the domestic benchmark concentration and the EU portfolio diversification rules.In summary, our study is the first to evaluate the consequences of both the assorted characteristics of domestic Eurozone benchmarks and the European regulation that prevents portfolio concentration (UCITS IV) in the appropriate estimation of AS. Furthermore, our study identifies truly active management in domestic equity funds in the Eurozone markets. This chapter has important implications for policymakers and practitioners of the domestic equity funds in the Eurozone. In the strongly regulated European markets, our unbiased approach allows both investors and market supervisors to identify the accurate levels of active management of each industry after considering both the regulation of portfolio diversification and the concentrated domestic equity benchmarks. Market supervisors will have a better picture of the active management map to develop appropriate regulations for the mutual fund industry. In addition, our approach should help practitioners and investors to effectively find out the level of active management of domestic equity funds and therefore provide information for fund management companies to replace actual performing managers. Further, our results should help to reduce the opacity in the management fees that funds charge by providing accurate measures of active management. In Chapter 2, we analysed how some market and fund characteristics play a crucial role in explaining the portfolio concentration default on Directive 2009/65/EC (UCITS IV). This chapter should help market supervisors to improve the monitoring process of defaults by domestic equity funds in the Eurozone mutual fund industries. On the one hand, the origin of UCITS directives was considered the beginning towards market protection and increased transparency in the Eurozone. On the other hand, the structure of mutual funds allows retail investors to access sophisticated active strategies that comply with liquidity and transparency restrictions protected by regulatory oversight. Their rules are based on a certain degree of portfolio diversification with the goal of reducing their vulnerability to portfolio risks. This chapter is the first to analyse how concentrated strategies could lead to non-compliance with article 52 in UCITS IV. First, we analysed several market characteristics that may influence the probability of a fund manager failing to meet the portfolio concentration limits. Using a logic panel data model (fixed effects), we estimated the probability of incurring defaults. Our findings provide evidence that should lead market supervisors to pay attention to concentrated fund industries with concentrated domestic benchmarks to prevent defaults on Eurozone concentration limit. The level of concentration of the domestic equity benchmarks would make the defaults almost twice as likely to occur. The level of concentration of the domestic fund industry also has positive and significant effects on the likelihood of incurring defaults. That is, defaults are approximately 12% to 17% more likely to occur when the level of concentration in the domestic fund industry increases. This evidence is consistent with the findings in the literature that link competition with active management strategies such as concentrated portfolios (see Dyck et al., 2013). Second, in the same line of Chapter 1, we analysed several funds characteristics that may influence the probability of a fund manager failing to meet the portfolio concentration limits. In Eurozone markets, market supervisors should especially monitor the most experienced funds that are solo-managed to prevent portfolio weights over the 10% limit. This finding is in accordance with Goldman et al., (2016) who shows that individual managers have much more concentrated portfolios than management teams. The fund’s age has a positive and significant influence on the likelihood of defaults. That is, the probability of incurring defaults is approximately 18% to 26% higher among older funds. Thus, this evidence is in accordance with the literature that links older funds with higher levels of idiosyncratic risk as a consequence of having more concentrated portfolios (see Amihud and Goyenko, 2013). Further, we analysed the characteristics of those stocks that were especially subject to more concentrated strategies and, therefore, more vulnerable to mutual funds’ investment policies. For that, we applied a multinomial logic panel data model (fixed effects) and found that the weight of the stocks in their benchmarks had a positive and significant effect on default on the EU portfolio concentration limit. The results show how the probability of a stock being subject to default is approximately 8% to 25% higher when the stock weight in the domestic benchmark increases. Thus, market supervisors should monitor stocks with large weights in domestic equity benchmarks.In addition, we also find how the stocks of domestic benchmarks that have been held longer are likely to be subject to concentration defaults. That is, defaults are approximately 13% to 24% more likely to occur when choosing stocks that belong to the benchmarks during the last 24 months. Thus, this result is consistent with the literature that argues the local advantage of reducing information asymmetry problems.Thus, we find how stocks that present low volatility have a greater likelihood of being subject to non-compliance with EU portfolio concentration limits. In terms of percentages, the probability approximately lowers between 19% to 27% when stock volatility increases. This finding is in accordance with the commitment to controlled risk strategies (see Huang et al., 2011). Therefore, market supervisors should pay more attention to these stock characteristics to monitor stocks that are more frequently overweighted above EU concentration limits. In summary, this chapter is the first to both analyse and identify the determinants of domestic equity funds’ failure to comply with the portfolio concentration limits of EU Directive 2009/65/EC. Furthermore, our study also determines the characteristics of the stocks subject to these non-compliant portfolios in domestic equity funds in the Eurozone. This chapter has important implications for market supervisors and policymakers in the mutual fund industries of the Eurozone. Our approach allows market supervisors with limited resources to identify and control non-compliant domestic equity funds by monitoring only some fund-specific characteristics. The improvement of this monitoring process should contribute to the financial stability of the EU asset management industry in terms of investor protection and market transparency. That is, mutual fund unitholders should be completely certain that their money is allocated to portfolios fulfilling the concentration limits required by the EU. Our findings also show a tool to assist EU market supervisors in identifying some explanatory mechanisms for stock weights that are over the EU concentration limits. Thus, our results may help supervisors identify what kind of domestic equity funds are more inclined to default and what kind of stocks are likely overweighted by these funds. Market supervisors could especially monitor these stocks to verify that domestic equity funds are meeting the concentration limits. Market supervisors should focus their limited resources on these types of stocks held by domestic equity funds to prevent defaults in portfolio concentration. Finally, our approach could also help retail investors control their risk profiles in terms of exceeding portfolio concentration limits. This application is in line with the reinforcement of investor protection of portfolio concentration. Investors could be sure that domestic equity funds fully follow the diversification requirements and market transparency provided by the UCITS directives.Chapter 3 presents a new perspective in the analysis of active management. We introduced dynamism by proposing a new version of AS that considers how the managers deviate their portfolios in two consecutive periods. We define dynamic Active Share (dAS) as a measure to capture over time the actual level of activity by comparing the differences against the benchmark in two consecutive periods. Our measure captures not only the long and short static positions in each stock included in the benchmark but also the previous long (short) positions that have been overweighted (underweighted) in the next period. Thus, dAS allows us to divide between investment decisions driven to spread portfolio weights closer to the benchmark (i.e., decisions that lead to a lower differentiation to the benchmark) and the other which is further from the benchmark (i.e., decisions that lead to a higher differentiation). Focusing on the most relevant fund industries in the Eurozone, we analysed the predictive power of AS first. The best results were found in the Spanish industry and, with less robustness, in Germany, the Netherlands, Austria, and Belgium. France only shows this relationship in the long term when we consider the CAPM alpha; while Italy, Finland, Portugal, and Greece fail to offer significant results or even present negative relationships. These results indicate that the prediction ability of AS presents assorted results as this relation is less clear than that presented in the literature (Cremers and Petajisto, 2009) Second, we examined the influence on the prediction power for performance of AS by splitting stocks in the portfolios of our sample into benchmark and non-benchmark . The results allow us to identify how the high proportion of prediction ability for performance of AS is explained by the investment in non-benchmark stocks. Although the AS’s share of overweight or underweight benchmark stocks is really related to stock picking ability, investing in overweighted stocks may mean distortions in their performance evaluations that can lead to spurious contributions to portfolio performance.Third, we applied and analysed the predictive power of dAS. The results provided evidence that the performance of non-benchmark stocks in high AS funds did not extend to mutual funds with high dAS. Thus, this measure is less sensitive to the weight of the portfolio that is invested in non-benchmark securities that is very relevant in Eurozone countries where domestic benchmarks are highly concentrated. The main advantage of the dAS compared to the AS is that our measure provides more information and can be split according to different investment decisions. For that reason, we proposed splitting up the dAS to examine which trading decisions add value to the portfolio. Our most interesting results showed that German funds presents a robust relationship between selling investment decisions and subsequently better performance. This relationship was also found for Dutch, Austrian, and Portuguese funds but with actual limited significance. This finding is in accordance with the level of activity showed by these industries in Chapter 1.Further, we analysed the contributions to dAS that was generated by those trading decisions that might be considered as managers’ bets. These bets could be those decisions that increased holdings that were already overweighted (buy bets) or decrease those holdings that were already underweighted (sell bets). Our findings show as in Germany and in some cases in the Austrian and Portuguese industries that there is subsequent significantly better performance in buying decisions. Focusing on selling decisions, only Finnish funds show positive and strong beliefs and performance. The results for predicting performance show that those portfolios with a higher concentration of these bets offer subsequent abnormal returns as this prediction ability is even higher than the seminal AS in some of the markets in this study.Accordingly, the results shown in this chapter are an interesting addition to Cohen et al. (2010) who try to identify the trades in which the managers have more confidence. The empirical findings show that those mutual funds with trading decisions with a stronger belief from the manager (i.e., decisions that lead them to deviate even more from the benchmark) outperform the remaining funds, especially when buying decisions are considered. This evidence is consistent with Karoui and Patel (2020) who show that the benefit of AS lies in the selection decision rather than the weighting decision.In summary, Chapter 3 is the first to introduce a dynamic perspective on AS to capture managers’ activity and skill. This chapter is relevant for investors who should be interested in knowing whether their fund manager is active and whether their decisions on new investment opportunities add value to the portfolio (i.e., undervalued assets) or on the contrary, the fund manager is passive. The common argument is related to the fees because the management fees charged by these two types of funds should be different. In addition, the chapter is also relevant to regulators in order to adjust the management fees charged by management companies to the actual level of activity carried out by mutual funds. Several fields for further investigation have been identified in this dissertation. Regarding the EU Directive, an extensive test of the level of enforcement that the UCITS IV has implemented in each Eurozone country (including emerging markets) could be useful for making decisions about possible changes in the regulation. Furthermore, a complementary analysis of chapter VII in UCITS IV in terms of its appropriateness with the characteristics of each Eurozone country could allow a more complete assessment of the limits of portfolio concentration and the implementation of the protection to investors. In addition, this dissertation makes an important advance in the topic of active management to help further an analysis in other areas or regions with similar characteristics that could be relevant to the financial literature.