It’s all in the name: The effect of mutual fund name ...



It’s all in the name: Mutual Fund Name Changes after SEC Rule 35d-1Susanne Espenlaub1(Susanne.espenlaub@manchester.ac.uk)Imtiaz ul Haq2(Imtiaz.haq@lums.edu.pk)Arif Khurshed1*(Arif.khurshed@manchester.ac.uk)ABSTRACTWe study how investors respond to ‘superficial’ mutual-fund name changes that occur for no fundamental reasons. We find that such name changes remain widespread even after regulation to curb potentially misleading name changes (SEC Rule 35d-1). Superficial changes are more widespread than previously studied ‘misleading’ changes that are not accompanied by corresponding portfolio adjustments reflecting the investment style suggested by the new name. Superficial changes appear to be driven by managerial incentives. Investors react to superficial changes with increased fund flows but appear to gain no benefit through improved performance or lower fees. On the contrary, name-change funds underperform as a group. Our findings highlight inefficiencies in the mutual-fund market and hold important implications for the stakeholders involved. JEL classification: G11, G14, G18Keywords: mutual funds; SEC Rule 35d-1; name changes; fund flows____________________________________________________________________1Alliance Manchester Business School, Booth Street West, University of Manchester, UK. Tel.: +44 161 27544752Lahore University of Management Sciences, Lahore 54792, Pakistan, Tel: +92 42 35608419* Corresponding authorWe would like to thank the editor Geert Bekaert and an anonymous referee for their useful suggestions. We are also grateful to the participants of the IFABS conference (Lisbon 2014), Krishna Paudyal, Brahim Saadouni and other colleagues at AMBS and LUMS for their insightful comments and suggestions. All remaining errors are our own.IntroductionIn 2001, the US Securities and Exchange Commission (SEC) introduced Rule 35d-1 to regulate mutual fund names. The rule aimed to address concerns that investors may be misled by fund names that do not reflect their true underlying asset holdings and risks. From 31 July 2001 onwards, Rule 35d-1 requires funds to invest at least 80% of their portfolios in assets suggested by the fund name. The regulatory intervention followed a significant number of fund name changes without corresponding changes in the underlying fund portfolios (Cooper et al., 2005). The issue was highlighted by practitioners in the 1990s: “An investment in Fidelity's Blue Chip Growth Fund certainly sounds like a stake in America's industrial giants. So shareholders might be startled to learn that their five-year return of 115 percent was achieved by a shift into risky medium-sized companies and international stocks. … The truth is, potato-chip makers are subject to stricter labeling requirements than mutual funds are - and that's a problem for investors faced with more than 7,000 stock and bond-fund choices.” (Source: U.S. News & World Report 1997)Information about a fund’s investments and related risks cannot be captured by the name alone, but the fund name can signal an association with a broad category of investment style. The SEC has recognized the importance of such communication in regulating fund names. In an industry with an increasingly diverse range of products, Rule 35d-1 is meant to reduce investor confusion and aid investors in making appropriate investment decisions. The controversy over fund names and the adoption of Rule 35d-1 increased the regulatory (compliance) costs to funds of changing their names, which should have reduced funds’ propensity to change names, all else equal. Moreover, one might expect that SEC warnings and Rule 35d-1 made investors more cautious of funds that change their names. As a result, it seems reasonable to expect that Rule 35d-1 would reduce the expected benefits to funds of name changes through abnormal fund inflows by misled investors. If the expected (regulatory) costs to funds of name changes increase, while their expected benefits fall, we would expect to observe comparatively fewer name changes after the adoption Rule 35d-1. While a name change can be useful to communicate a change in some underlying fundamentals of the fund (e.g. a portfolio change), we should expect managers to be more conservative following the new regulation, particularly for name changes that are not motivated by any such fundamental changes.Our study examines all name changes by US equity funds after the adoption of Rule 35d-1 during 2002(Q4)-2011. Our sample period starts after the publication of Rule 35d-1, and after its compliance date in September 2002, by which time funds with previously misleading names had to have adopted more appropriate names. To our knowledge, this is the first study of mutual fund name changes after the adoption of Rule 35d-1. We have a relatively large sample size of 2,677 name changes compared to previous studies, which additionally allows us to explore some of the heterogeneity issues across the fund industry in the context of name changes.Our study is related to Cooper, Gulen and Rau (2005) who analyse funds that adopt new names that suggest specific investment styles. Based on data before the implementation of Rule 35d-1, Cooper et al. find that name changes give rise to significant abnormal inflows into funds that change their names, particularly for funds that are renamed to reflect recent winning investment categories (‘hot styles’). They find that 63% of style-related name changes are ‘misleading’ in that they are not accompanied by corresponding changes in portfolio holdings to reflect the investment style suggested by the new name. We find that misleading name changes continue to occur even after the introduction of stricter regulation through Rule 35d-1, although the proportion is lower at 20% of style-related name changes in our sample. Following the introduction of Rule 35d-1 and the SEC advice to investors not to base investment decisions on fund names alone, we expect investors to be more wary of fund-name changes and exercise greater caution when investing in funds following name changes. Hence, we would expect investors in the market to treat misleading name changes differently from non-misleading ones. There is no reason to expect abnormal fund inflows in response to the former. Contrary to expectations, we find significant abnormal flows into funds that make misleading name changes. The magnitude of these abnormal flows in the first quarter after the event is comparable to that found by Cooper et al. (2005), although we do not find a substantial long-term affect. Interestingly, we do not find evidence of negative drift in abnormal flows for up to one year post event. It is worth noting that while Cooper et al.’s sample consists of only 332 name changes around certain investment styles, our results are based on a larger, more inclusive sample (2,677 changes) that includes all types of name changes in equity mutual funds.We analyse determinants of short-term and long-term abnormal flows to better understand the drivers of such behaviour. Even after controlling for various factors (including marketing and hot styles), we find investors do not differentiate in their reaction between misleading and non-misleading name changes. The non-differential response is seemingly in contrast to expectations following the Rule 35d-1, as investors appear to treat both types of name changes as a positive signal. This investor reaction may be a rational response after all. While a misleading name change may not have a corresponding portfolio change, it may be accompanied by a fundamental change in some other driver of future earnings potential of the fund (such as a change in ownership or management). In such a case, a name change labelled as ‘misleading’ may very well hold information that would be useful to an investor. This leads us to ask a broader question with potentially wider implications: Do investors react differently when name changes are accompanied by fundamental changes to the fund to when they are not? We address this question by studying investor reaction to name changes that we classify as ‘superficial’ in the sense that they lack a fundamental cause. More precisely, these are name change events not warranted by changes in a fund’s investment strategy or holdings, a merger or acquisition, or a change in the fund manager. We document widespread adoption of such name changes by mutual fund managers; in fact, these constitute a majority (approximately 60%) of all the name changes in our sample period. We further find that such changes are more likely to occur as a response to lower managerial compensation as a result of a shrinking asset base over time and lower management fees (in percentage terms of asset base).We should not expect investors to react to non-fundamental, superficial name changes as these events contain absolutely no information regarding a change in the future potential earnings of the fund. In contrast to our expectations, however, we find that not only do such name changes attract significantly positive abnormal flows but that these are not significantly different than those of fundamental, non-superficial name changes. This is indicative of a much larger inefficiency in the market than misleading name changes; a superficial name change is approximately four times more likely to occur than a misleading name change (there are only 435 misleading changes in our sample period as opposed to 1,605 superficial changes). Interestingly, the magnitude of abnormal flows to superficial name changes is also larger than that to misleading ones. As before, controlling for other factors does not change our findings. We also find no evidence that performance improves after such name changes or that fees decrease, indicating some irrationality on part of investors. Instead, we find that investors tend to lose out due to lower returns to name-change funds as a group.The rest of the paper is structured as follows. Section 2 reviews the relevant SEC regulation. Section 3 describes our data and methodology. Section 4 outlines and discusses our results. Section 5 concludes the paper.2. Institutional Background on SEC RegulationIn March 2001, the Securities and Exchange Commission (SEC) introduced a new rule under the Investment Company Act of 1940. The new legislation, known as Rule 35d-1, was adopted to regulate certain broad categories of investment company names that are likely to mislead investors about the fund’s investments and risks. It requires a mutual fund to hold at least 80% of its portfolio in the type of security indicated by its name. In an industry with an increasingly diverse range of products, this rule is meant to reduce investor confusion and increase investor confidence. Regulating against misleading names is fraught with conceptual difficulties, primarily the need to identify instances of misleading names. Rule-based regulation requires a comprehensive list of names and associated acceptable (as opposed to misleading) investment holdings and strategies. The notes and comments contained in the legislation document published by the SEC in 2001 recognise and discuss these issues. While the rule enforces strict conditions on names that suggest a particular type of holding, it explicitly exempts a number of fund names that suggest types of investment styles (although their adoption was to be monitored on a case-by-case basis by the SEC). While these exemptions are ostensibly due to the conceptual difficulties in defining investment styles, they are likely to create loopholes that may allow mutual funds to take advantage of (uninformed) investors. According to the legislation, names that include ‘value’, ‘growth’ and ‘balanced’ are not subject to the new legislation. The use of the word ‘foreign’ is regulated since it indicates investments tied economically to countries outside the U.S. However, funds are free to use ‘global’ and ‘international’ as these “do not suggest a particular investment focus, but rather a particular type of diversification among different investments” according to the SEC. Appendix A provides more information on the main categories of name changes which are subject to Rule 35d-1.3. Data and MethodologyThe data used in this study is sourced from the Centre for Research in Securities Prices (CRSP) mutual fund database. We use quarterly data since fund name changes are tracked at this frequency. Our sample period starts from the fourth quarter of 2002, after the SEC compliance date for Rule 35d-1 has passed (July 31, 2002), to ignore the effect of funds changing names simply for compliance reasons. Following previous literature in mutual funds, we limit ourselves to only US equity mutual funds. We keep track of name changes in our dataset for which the CRSP fund number, a permanent unique identifier for each share class, does not change. This results in 5,078 name change events once administrative name changes are excluded. We then drop events for funds with fewer than 12 quarters of data, leaving 4,666 changes. Finally, we control for multiple share classes to avoid duplicate counting by aggregating across share classes. Our final sample contains a total of 2,677 name changes during Q4/2002 to Q4/2011. The number of name-change funds in our sample is lower (2,110) as some funds change name more than once. There are on average 3,616 domestic equity mutual funds in the U.S. market per year during our sample period and approximately 8% of these undergo name changes every year. We categorize name changes along two dimensions; misleading and superficial. For the former, we label a name change as misleading if the new name is not reflective of the fund’s underlying portfolio (based on its Lipper classification), and non-misleading otherwise. For the latter, we label a name change as superficial if it is not associated with changes in a fund’s investment strategy or holdings, a merger or acquisition, or a change in the fund manager, and non-superficial otherwise (see Appendix B for more details). Superficial name changes are much more popular in our sample, making up about 60% of all the changes. Name changes are not distributed evenly over time. Figure 1 plots the number of name changes per year in our sample period. There is a higher occurrence in changes in years around the financial crisis of 2007-8. The year 2010 witnesses an unusually high number of name changes. This may have been driven in part by the high volatility in equity markets and in part by the introduction, in the same year, of the Dodd-Frank Wall Street Reform and Consumer Protection Act. Sections 918 and 919A of the Act required the regulatory body to investigate the information mutual funds share with investors for the purposes of improving investor protection, which may have prompted funds to undertake a greater number of changes in anticipation of the increased scrutiny. Figure 1 also reports the number of misleading and superficial name changes per year. INSERT FIGURE 1 HEREIn the absence of data on actual inflows and outflows to mutual funds, we follow the method of Sirri and Tufano (1998) to compute net flows. The normalized fund flow for quarter t is defined as: TNAt- TNAt-11+rt- MGTNAtTNAt-1(1)where TNAt is the fund’s total net assets at the end of quarter t, rt is the fund’s return over quarter t, and MGTNAt is the increase in a fund’s asset base due to a merger or acquisition in quarter t. We assume fund flows to occur at the end of every quarter using this formula. A summary of the variable definitions can be found in Appendix C. 4. Analysis and Results4.1. Descriptive StatisticsTable 1 compares the descriptive statistics of the name-change funds against the universe of all other equity mutual funds, aligned on the date of the name change. Panel A of Table 1 shows that name-change funds have relatively lower total net assets and net fund flows prior to changing their name. Non-change funds have higher lagged excess returns than name-change funds. The difference persists despite controlling for risk, although the four-factor alphas are negative for both groups. This is, however, not the case for one-year compounded excess returns. Mean quarterly flows over the previous year continue to be different for the two groups. Lagged 12b-1 fees are higher for name-change funds while there is little difference between the lagged management expenses.Panel B confirms the differences reported in Panel A. The panel shows how fund characteristics for name-change funds are distributed relative to other funds. Using a χ2 one-sample test for uniform distribution across five quintiles, we find the distribution of all variables in Table 1 is significantly different across the two groups. Name-change funds are smaller (as measured by TNA), face lower net flows (in previous quarter and year) relative to other funds and tend to underperform their counterparts (in terms of both raw and risk-adjusted returns). There is dispersion towards the tails for management fees while a majority of the funds fall in the largest two quintiles for 12b-1 fees. Funds that change names also tend to be younger than their counterparts.INSERT TABLE 1 HERE4.2. Determinants of Mutual Fund Name ChangesWe carry out a more formal test in this section on the types of mutual funds that change their names and the point in their lifecycle at which they choose to do so. Table 2 shows the relationships between main fund characteristics and the likelihood of a fund changing its name. Panel A employs a logistic regression to examine name change determinants. There are 37 cross-sections pooled together, one for each quarter in the sample. For each of these, the dependent variable is a dummy that is assigned a value of 1 for a fund that changes names that quarter and 0 otherwise. The independent variables are lagged returns, lagged fund flows, log of lagged total net assets (TNA), lagged management expense, lagged 12b-1 fee, lagged four-factor alpha, mean fund returns and mean fund flow in past 4 quarters and log of fund age (in quarters). Different columns refer to different categories of name changes. The first column includes all name changes. The second column is a logistic regression where the dependent variable adopts a value of 1 for a misleading name change and 0 otherwise. This changes to a value of 1 for non-misleading names, superficial and non-superficial for columns 3, 4 and 5 respectively. We find that marketing (12b-1) fees is a significant factor across all columns. This is in line with our expectations; greater marketing expenditure is associated with higher probability of all types of name changes. A name change is only effective if it is communicated properly to the market. As we show later, there is little evidence of an increase in this expenditure over the event. Hence, marketing expenses prior to the name change event seem to act as a pre-requisite. We also find that younger funds are more likely to undertake different types of name changes, with the sole exception of misleading changes.A major difference between misleading and non-misleading name changes, as well as superficial and non-superficial, is the role of fund performance. We find that both raw returns and risk-adjusted returns are related to the probability of the non-superficial name changes. This is in line with our expectations as a change in a fund’s fundamentals is essentially an attempt to improve future performance (which is also the information contained in such a name change) and hence more likely a response to dismal past performance. In other words, the poorer the past performance, the more likely a fund will attempt to change its strategy, portfolio, ownership or management and hence more likely it will engage in a non-superficial name change. This is not the case with superficial name changes, which explains why we do not find any significant connection to prior fund performance. The same argument extends to misleading and non-misleading name changes. A second major difference between these types of name changes is the role of management fee prior to the event. Lower management fees are associated with a higher probability of both superficial and misleading changes. Management fees are structured as a percentage of the assets under management. With lower fees, there is greater pressure on managers to increase asset size in order to maintain personal incentives. This could lead to a greater tendency to engage in tactics that increase the asset base of the fund. A superficial or misleading change is one low-cost strategy in this case. Hence, lower managerial incentives may lead to greater probability of such name changes. This argument is less relevant for non-superficial or non-misleading changes as the costs involved are higher due to a change in the underlying portfolio, strategy, ownership or management. Our results seem to support this notion. We investigate this further by carrying out a time-series logistic regression for the sample of funds that change names to examine when funds choose to change their names. The results are reported in Panel B. We regress a dummy variable that takes the value of 1 in the name-change quarter and 0 in all other quarters on the lagged explanatory variables. We use all the time-series data available for each name-change fund. We report results on the same categories of name changes as in panel A.Panel B corroborates our earlier results. We fund that performance is relevant only for non-superficial and non-misleading name changes, although only raw returns are significant here. The four-factor alpha tends not be significantly different before a name change event than at other points in a fund’s lifecycle. The negative relation with raw returns is in line with greater name changes observed during periods of market downturn. As before, we find no connection between fund performance and superficial or misleading name changes. A comparatively smaller asset base during a fund’s life seems to be a common driver amongst all types of name changes. Lagged flows during the previous quarter and year are not significant, suggesting the smaller asset base is not the result of recent outflows. As before, funds are more likely to change names (except for misleading changes) earlier in their life.INSERT TABLE 2 HERE4.3. Impact of name changes on abnormal fund flowsOur results so far suggest that managers concerned with poor performance are more likely to shift to a different investment strategy or market, change their investment portfolio, merge or acquire another fund and/or change their management, all of which are normally followed by a non-superficial or misleading name change. On the other hand, managers more concerned with personal incentives seem to be simply undertaking superficial or misleading name changes, a strategy that does little to affect performance but may be successful in attracting flows if it is treated as a positive event by investors. In this section, we investigate whether this is an effective strategy on the part of the fund manager. We do so by analysing fund flows around the name change event. The main issue in this approach lies in the endogeneity that arises due to the fact that changing names is not random but a deliberate decision undertaken by fund managers, who self-select into their preferred choices. To control for this non-random selection bias, we adopt the endogenous switching regression model, an extension of the two-stage Heckman (1979) model. We choose this model over the propensity scoring matching (PSM) method to account for unobservables in addition to observable fund characteristics. This methodology is better suited to deal with situations where the decision to choose treatment is an endogenous choice of the decision-maker, as in the case of a fund manager choosing to change fund names (Li and Prabhala, 2007). We choose the extension over the baseline Heckman model for the purpose of constructing a counterfactual analysis for comparison purposes and hence producing treatment effects for easy interpretation.We follow the method of Maddala (1991). The selection model for the first step of the model is as follows:INC =γ0 + γ1.LnTNA + γ2.LagFlow + γ3.LagRtn + γ4.LagAlp + γ5.Lag12b + γ6.Lagexp + γ7.LagMeanRtn + γ8.LagMeanFlow + γ9.LnAge + Quarter dummies + μ (2)where INC is a dichotomous variable that equals 1 if a fund changes name and 0 otherwise, LnTNA is the log of the lagged total net assets (in $millions), LagFlow is the lagged quarterly flow, LagRtn is the lagged quarterly excess returns, LagAlp is the lagged quarterly four-factor alpha, Lg12b is the lagged marketing expenses (12b-1 fees), Lagexp is the lagged management expense ratio, LagMeanRtn is the mean quarterly fund returns over the past four quarters, LagMeanFlow is the mean quarterly fund flow over the past four quarters, LnAge is the log of the fund age in quarters, and μ is the stochastic disturbance term. This binary outcome equation determines the decision to change names and is estimated using a probit model. It provides the two inverse Mill’s ratios required in the next step.The two equations for the outcome model in the second stage are:Flow1 =β0 + β1.LagRtn + β2.LagAlp + β3.Lag12b + β4.Lagexp+ β5.Lag12b*LagRtn + β6.Lag12b*LagAlp + β7IMR1 + ?1(3)Flow0 =δ0 + δ1.LagRtn + δ2.LagAlp + δ3.Lag12b + δ4.Lagexp + δ5.Lag12b*LagRtn + δ6.Lag12b*LagAlp + δ7.IMR0+ ?0(4)where LagRtn is the lagged excess returns, LagAlp is the lagged four-factor alpha, Lag12b is the lagged marketing expenses (12b-1 fees), Lagexp is the lagged management expense ratio, Lag12b*LagRtn is the interactive term between lagged 12b-1 fees and lagged excess returns, Lag12b*LagAlp is the interactive term between lagged 12b-1 fees and lagged four-factor alpha, IMR is the respective inverse Mill’s ratio and ? is the respective error term. All lagged variables refer to the quarter prior to the name change. Flow1 refers to the flow to funds that undergo a name change while Flow0 refers to the alternative regime which comprises of all other funds. The choice for the explanatory variables in the outcome model is influenced from past literature on the topic. Several previous papers (Chevalier and Ellison, 1997; Wilcox, 2003; Patro, 2006; Cashman et al., 2011) find that past performance influences flows to funds. Hence we include past excess returns and the risk-adjusted returns. In addition, Barber et al. (2005) also find that funds that spend more money on advertising and promotion (i.e. those with higher 12b-1 fees) are found to attract greater flows. The inclusion of the lagged 12b-1 fee denotes that minimization of search costs on part of the investors is an important factor in their decision making process. Meanwhile, Sirri and Tufano (1998) report that funds which advertise their recent superior performance are able to capture substantially more flows than their counterparts. Therefore, we also include interaction terms between 12b-1 fees and past excess returns, as well as between 12b-1 fees and prior four-factor alpha. Lastly, we include the lagged management expense ratio as Ivkovic and Weisbenner (2009) argue that this variable acts as a signal for managerial talent in the investment decision. In order to observe the immediate and long-term impact of the name change, we carry out the second stage with two alternative dependent variables: normalized flows for the quarter after the name change and normalized flows for the year after the name change. The inverse Mill’s ratio is significant for both models in the second stage, confirming that unobservables do contribute to the self-selection bias and hence should not be ignored.We then use the estimates from the second stage to conduct a counterfactual analysis on alternative scenarios that are unobserved. For example, we are able to estimate the flows to a name-change fund if it had instead not chosen to change names. Hence, we are able to compute abnormal flows arising due to a name change by simply subtracting the counterfactual estimated flows from the actual realized flows:Abnormal flows for fund i = Flow1,i – Flow0,i = Flow1,i – (X1,i . δ0) (5)The counterfactual estimates are obtained by multiplying the estimated parameters δ from equation (4) with the realized values of the exogenous variables of funds which undergo a name change (from equation (3)). The resulting series is the expected flows for name change funds in the counterfactual scenario as if they had not undertaken a name change. We use these to compute abnormal flows after the name change event. The above procedure essentially produces treatment effects by isolating the variation in fund flows that arises solely from a fund changing its name. Panel A of Table 3 reports the abnormal flows calculated for different types of name change funds by subtracting the computed counterfactual flows from the observed flows, as per equation (5). The short-term impact is measured by abnormal flows in the quarter immediately subsequent to the name change while the longer-term impact is depicted by the mean abnormal flow per quarter over the entire year post event. We find that name changes result in significantly positive abnormal flows. Approximately $59 million of abnormal flows result to each name-change fund on average in the subsequent year due to the event itself. Collectively, this accounts for 23.8% of the average total flows in the domestic equity sector per year in our sample period. Interestingly, both misleading and non-misleading changes attract abnormal flows. One can expect this result for name changes that are undertaken along with a corresponding change in the fund’s portfolio as this event may result in a positive signal due to information regarding a possible change in the future earning potential of the fund. Hence, a non-misleading change can result in positive abnormal flows in an efficient market. However, this should not be the case for non-misleading name changes as there is no change in the holdings. Yet we document that such changes witness an average of 8.32% abnormal flows over the subsequent year, with most of it arising in the first quarter. These results are in line with those of Cooper et al. (2005), who also find that investors react positively to such changes, although their results suggest a long-term trend. This response appears irrational and so prompts us to ask a broader question with potentially wider implications: Are investors influenced by purely superficial name changes?A closer inspection reveals that misleading names may not be purely superficial in nature. While a misleading name change may not have a corresponding portfolio change, it may be accompanied by a change in some other driver of future earning potential of the fund (such as a change in ownership, management or strategy). In such a case, a name change labelled ‘misleading’ may very well hold information about future returns that would be useful to an investor. Hence, one would expect investors, even in an efficient market, to react to such changes by diverting more money towards these funds (assuming the signal is positive). To observe investor reaction to such non-fundamental name changes, we turn to the superficial category of name changes. We know, based on our construction of this category, that there is no information contained in such name changes and hence we would not expect rational investors to either increase or decrease their flows to such funds after a name change. However, panel A reports that even superficial name changes attract significantly positive abnormal flows. Remarkably, these are on average greater in magnitude than those by misleading name changes, especially over the subsequent four quarters. Panel B of Table 3 tests for differences in cumulative abnormal flows between misleading and non-misleading, and superficial and non-superficial. While there is no difference in the subsequent quarter, misleading name changes in the longer run appear to earn significantly less than their counterparts. We explore this further in the next section. Interestingly, such is not the case between superficial and non-superficial, where we do not find any significant difference in investor reactions in both the short and long term. This offers some evidence of irrationality on investors’ part. It also explains the high prevalence of superficial name changes in our sample period; such name changes do seem to be effective in influencing investors and hence form a rational strategy for fund managers looking to attract greater flows. The introduction of new legislation and SEC’s repeated warnings does not appear to have been successful in protecting investors from such deceptive practices surrounding mutual fund name changes. INSERT TABLE 3 HERE4.4. Determinants of abnormal fund flowsWe investigate the seemingly unexplained flows of funds to name changes, especially superficial ones, by examining the determinants of abnormal flows. It is important to control for variables, such as marketing, that may explain why certain funds attract more abnormal flows despite a superficial name change. In doing so, we also test Cooper et al.’s (2005) theory that investors are chasing ‘hot style’ name changes, regardless of whether these are fundamental or not. We follow their methodology in defining hot style names; a move towards (away from) a name reflecting a particular investment style when the associated style premium is positive (negative) over the last two quarters is classified as ‘hot style’ name change. In addition to this, we also test how investors treat changes where the new name exploits one of the explicit loopholes identified in the SEC regulation itself (which is public information as mentioned in section 2). Essentially, these are names that include the terms ‘value’, ‘growth’, ‘balanced’, ‘international’ or ‘global’.We run an OLS regression with the cumulative abnormal flows in the quarter after the name change as our dependent variable in the following equation:Abnormal Flow+1q =γ0 + γ1.Superficial + γ2.Misleading + γ3.HotStyle + γ4.Loophole + γ5.LnTNA + γ6.LagFlow + γ7.LagRtn + γ8.LagAlp + γ9.LnAge + γ10.?12b + γ11.?exp +Quarter dummies + μ(6)where Superficial is a dummy that equals 1 if the name change is superficial and 0 otherwise, Misleading is a dummy that equals 1 if the name change is misleading and 0 otherwise, HotStyle is a dummy that equals 1 if the name change is a hot style name change and 0 otherwise, Loophole is a dummy that equals 1 if the name change exploits a loophole in the new regulation and 0 otherwise, LnTNA is the log of the lagged total net assets ($ millions), LagFlow is the lagged quarterly fund flow, LagRtn is the excess return in the quarter prior to the name change, LagAlp is the lagged quarterly four-factor alpha (calculated over quarters -12 to -1 before the name change), LnAge is the log of the fund age in quarters, ?12b is the change in the annual marketing expense (12b-1 fees) in the quarter before and after the event, ?exp is the change in the annual management expense in the quarter prior and post the name change, Quarter dummies are time dummies for each quarter and μ is the stochastic disturbance term.Panel A of Table 4 reports the results. Even in our basic model (model 1), we do not find that investors differentiate between superficial and non-superficial name changes, as depicted by the insignificant beta on the superficial dummy. We do find that a significant and negative coefficient for fund size (log of TNA). Since the dependent variable is normalized abnormal flows, the negative sign could be due to smaller, lesser known funds finding it easier to move away from their past image to adopt a new one, as opposed to larger, more popular funds that have stronger prior association with investors. In model 2, we find that misleading name changes attract lower flows with the inclusion of some controls, although this is not significant. We find in model 3 that investors appear to value hot style name changes more than others, in line with Cooper et al. (2005). The loophole dummy in model 4 is significant and negative; it appears that, as expected, investors treat names exploiting explicit loopholes identified in the SEC regulation itself (and hence public information) negatively. The change in the p-value for the misleading dummy over the previous model suggests that at least some of misleading name changes tend to exploit loopholes. This also explains why we continue to observe, in our sample period, name changes suggesting certain investment styles without a corresponding portfolio change despite the new regulation attempting to clamp down on exactly such actions (although we note that not all misleading names use loopholes). However, once we include all the variables in the full model (6), we note that the loophole dummy becomes insignificant while change in management expense ratio is significantly negative. Investors seem to direct less flows towards funds that increase this expense over the event. It might be the case that many misleading changes exploiting loopholes also increase the management fee, as we would expect from managers who are trying to maximize personal incentives. The hot style dummy stays significant in all models, lending support to Cooper et al.’s (2005) findings.We repeat the analysis in Panel B for the longer term impact by changing our dependent variable in equation (6) to the mean quarterly abnormal flows over the entire year following the name change. Some of our results differ from Panel A. We find that the hot style dummy is not significant in explaining abnormal flows over the year, only in the short-run. We also find that the misleading dummy becomes insignificant only when the change in management expenses is included in the model. Furthermore, we find that the older the fund, the lower the abnormal flows. In addition to the above, we carry out several out robustness checks to test if our results are driven by other factors (results unreported). Firstly, we check whether flows are concentrated in certain individual styles. We individually test for investor reaction to changes where the new name includes the following ten most popular terms: ‘value’, ‘growth’, ‘small’, ‘mid’, ‘large’, ‘diversified’, ‘international’, ‘global’, ‘aggressive’, and ‘America(n)/U.S.’. We do so by separately conducting regressions on the full model in both Panel A and B of Table 4 where we include a dummy for each of the above terms, but do not find any significant evidence that investors actively invest or divest from such name changes, indicating that individual styles are irrelevant, beyond the categories we identified earlier.Secondly, we check whether abnormal flows are limited to either retail or institutional investors. In order to do so, we carry out an analysis without aggregating share classes and introduce in equation (6) a dummy that takes the value of 1 if the share class is exclusively available to retail investors and zero otherwise. We do not find that the effects are contained to either set of investors.Thirdly, we check whether abnormal flows are contained to funds nested within larger fund families. We are interested in distinguishing fund families on size since larger families may benefit from stronger branding and recall amongst investors, especially in the context of name changes. We divide families based on size into quartiles. We look at the difference in investor reaction to name changes amongst funds in different sized families in the full model in Table 4 but find no significant difference. In summary, we find that investors do not discriminate between superficial and non-superficial name changes (nor between misleading and non-misleading). The relevant superficial dummy is not significant and negative in all our models, implying that investors do not care about the superficial nature of the name change even when incorporating other factors into the equation. INSERT TABLE 4 HERE4.5. Changes in fund characteristics around name changesWhile we have established that investors are indeed reacting to superficial name changes, we have more to prove in order to determine that this is in fact a non-rational response. Although there is no reason to suspect that a superficial name change will contain any useful information, it might be the case that a simple increase in the fund’s assets under management has unrelated benefits for its investors. Two possibilities stand out; a decrease in expenses as a result of greater economies of scope due to the larger fund size, and an increase in returns for the fund as a result of directing inflows to existing holdings. Khorana & Servaes (2012) show that larger funds families have lower expenses as a percentage of the assets under management and that many of the funds pass on the savings to their investors through lower management fees. Hence investors may stand to benefit from a name change that is likely to pull in more flows. Regarding the latter, Wermers (1999, 2004) show that inflows that are invested in existing portfolios can drive up these stock prices, resulting in higher returns recorded for funds engaging in this practice. To check whether either of these takes place, we check for changes in fund performance and expenses around the name change event. Table 5 reports the change in average fund characteristics from the year before the name change to the year afterwards. We find no change in raw returns looking at the entire group of name changes. Upon segregation, we find a significant decrease in performance for both the superficial and misleading name changes. On the other hand, raw returns improve for their counterparts (results unreported). The four-factor returns prove that there is no significant difference in the risk-adjusted returns over name changes that are superficial and misleading. Once again, we find that performance for their counterparts does increase, suggesting that on average the fundamental changes undertaken are successful (results unreported). This may also be responsible for the favourable conditioning of investors towards name changes in general. As far as expenses are concerned, we find that management fees tend to increase rather than decrease for superficial name changes, contrary to our expectations. The increase, however, is small in magnitude but nevertheless dispels the economies of scale theory. Marketing (12b-1) expenses change only very marginally (0.007%) for superficial name changes although the difference is statistically significant.INSERT TABLE 5 HERE4.6. The impact of additional name changesOur results suggest that a superficial name change in its own is a simple strategy to lure investors. A natural question then arises: do such name changes remain as effective in attracting investor money over repeated uses? We are able to answer this question because there are funds in our sample period that change names multiple times. Approximately 20% of the changes in our dataset are multiple name changes. A majority (63%) of these are superficial in type. We repeat the analysis for Table 3 for additional name changes. Table 6 presents the abnormal flows calculated from the counterfactual analysis in equation (5). We find that additional name changes continue to attract significantly positive flows. However, these are significantly lower in magnitude than their counterparts for both the short and long term. Interestingly, we find that additional superficial name changes also attract significantly positive flows, although almost all the gains are contained in the first quarter after the event. As before, we find no evidence that investors differentiate between superficial and non-superficial name changes. However, we find that the magnitude of the abnormal flows that repeat superficial changes attract is significantly lower than that of non-repeat superficial changes, implying that at least some portion of investors exhibit learning over multiple events. From a fund manager’s perspective however, a superficial name change does not seem to be a strategy that is limited to a one time use.INSERT TABLE 6 HERE4.7. Discussion and implicationsOur paper has important implications for the main stakeholders of the mutual fund industry. From a fund manager’s perspective, we show that changing names can be a simple yet effective strategy to attract investor flows, one that can be used multiple times over a fund’s life. Our results show that the average fund earns an additional $0.55 million in annual fund fee revenue as a result of a name change, an increase of 13% over previous fees. While this is only 4% of the aggregate industry fees, it accounts for an astonishing 23.5% of all the fees earned on new flows to the industry per year. More interestingly, the manager is not required to undergo any corresponding fundamental change in the fund’s operations to capture these gains. While name changes are attractive to fund managers given that their incentive structure is based on the total assets under management, investors need to be wary of such changes, especially those that are purely superficial in nature. Our results show that investors are unlikely to benefit from a superficial name change, either through an improvement in fund performance or a reduction in fund fees. On the contrary, fund investors are expected to lose out given that superficial name-change funds underperform their counterparts as a group. In order to determine the loss to investors of their decision to direct more money towards superficial name-change funds, we look at the returns forgone had investors not been influenced by such changes. We do this by comparing the returns earned by the subset of superficial name-change funds to the overall industry returns. We choose the industry average returns as our reference point since there is weak evidence in literature on fund-selection ability by mutual fund investors (Frazzini & Lamont, 2008). Hence a random allocation across the industry serves as the benchmark for comparison. We find that the subset of superficial name-change funds significantly underperforms this benchmark. This drop in performance is shown in Appendix D: a decrease in average quarterly excess returns from 1.15% to 0.43% and a decrease in the average quarterly risk-adjusted returns (four-factor alpha) from -0.68% to -1.33%. Investor loss is not just restricted to superficial name changes: fund flows to name changes in general tend to underperform (by 0.89% in raw returns and 0.44% in risk-adjusted returns per quarter). However, returns to superficial changes are remarkably bad, especially in terms of risk-adjusted returns as they are responsible for 85% of the total wealth loss from name changes. Most of this loss is a result of poorer investment performance rather than higher fees (which only adds 0.10% to the loss per year).Clearly there are implications for policy-makers as well. Our results highlight significant investor losses due to switching to superficial name changes. This is not a small group of investors either; superficial name changes constitute a majority of name changes and account for approximately 14% of the total investments into mutual funds annually during our sample period. This translates into a significant loss on new investments, reducing raw returns by 8.7% and risk-adjusted returns by 13.2% on average on the aggregate industry inflows.5. ConclusionThis paper examines investors’ reaction to name changes by U.S. equity mutual funds following the adoption of Rule 35d-1 by the SEC in 2001. This regulation was meant to improve investor protection against questionable practices involving mutual fund names. Nevertheless, we document the occurrence of misleading and superficial name changes during our sample period. The latter is especially common, accounting for a majority (60%) of name changes. Such changes appear to be motivated by managerial incentives; lower fees exert greater pressure on managers to increase asset size to maintain compensation, which translates into a greater tendency to undergo a superficial name change.Contrary to our expectations, superficial name changes manage to attract significant and positive abnormal flows. Investors do not appear to be distinguishing in their response to fundamental and non-fundamental name changes. We find that our results do not change despite accounting for other factors (including marketing and hot styles). Neither do we find evidence that this phenomenon is contained to a certain class of investors, fund families or investment styles. Furthermore, we find no evidence that investors are benefiting from such name changes through either improved performance or decreased fund fees. Lastly, we document that additional superficial name changes continue to attract significantly positive abnormal flows, although these are lower in magnitude than their non-additional counterparts. The SEC Rule 35d-1 is meant to control the misuse of certain terms in fund names. Our results highlight concerns about not just misleading name changes, but a much broader category of superficial name changes that are successful in influencing investment decisions, at the cost of the investor. One of the issues is the inclusion of explicitly acceptable names that provide safe harbours, allowing funds to limit their expected costs of name changes. In fact, a majority of superficial name changes in our sample do exploit such loopholes. This appears to be a case of unintended side effects of the legislation. Overall, our results suggest the need for more direct regulation beyond Rule 35d-1 in order to improve investor protection in the mutual fund industry.ReferencesBarber, B. M., Odean, T., Zheng, L., 2005. Out of Sight, Out of Mind: The Effects of Expenses on Mutual Fund Flows. Journal of Business. 78, 381-14.Barberis,?N.,?Shleifer, A., 2003. Style investing. Journal of Financial Economics. 75, 283-317.Capon, N., Fitzsimons, G. J., Alan Prince, R., 1996. An individual level analysis of the mutual fund investment decision. Journal of Financial Services Research. 10, 59-82.Carhart, M. M., 1997. On persistence in mutual fund performance.?The Journal of finance.?52(1), 57-82.Cashman, G. D., Nardari, F., Deli, D. N., Villupuram, S. V., 2011. Investor behaviour in the mutual fund industry: evidence from gross flows. Journal of Economics and Finance. 10, 1-27.Chevalier, J., Ellison, G., 1997. Risk-Taking by Mutual Funds as a Response to Incentives. Journal of Political Economy. 105, 1167-1200.Cooper, M. J., Dimitrov, O., Rau, P. R., 2001. A by any other name. Journal of Finance. 56, 2371–2388.Cooper, M. J., Gulen, H., Rau, P. R., 2005. Changing names with style: Mutual fund name changes and their effects on fund flows. Journal of Finance. 60, 2825-2858.Cooper, M. J., Ajay, K., Igor, O., Patel, A. Rau, P. R., 2005. Managerial actions in response to a market downturn: Corporate name changes during the decline. Journal of Corporate Finance. 11, 319–335.Del Guercio, D., Tkac, P. A., 2008. Star power: The effect of Morningstar ratings on mutual fund flow. Journal of Financial and Quantitative Analysis. 43, 907-936.Frazzini, A., Lamont, O. A., 2008. Dumb money: Mutual fund flows and the cross-section of stock returns.?Journal of Financial Economics.?88(2), 299-322.Heckman, J. J., 1979. Sample selection bias as a specification error. Econometrica. 47, 153-161.Ippolito, R. A., 1992. Consumer reaction to measures of poor quality: Evidence from the mutual fund industry. Journal of Law and Economics. 35, 45-70.Ivkovi?, Z., Weisbenner, S., 2009. Individual investor mutual fund flows. Journal of Financial Economics. 92, 223-237.Keswani, A., Stolin, D., 2012. Investor reaction to mutual fund performance: Evidence from UK distribution channels. Journal of Financial Research. 35, 425-450.Khorana, A., Servaes, H., 2012. What drives market share in the mutual fund industry??Review of Finance?16(1), 81-113.Li, K., Prabhala, N. R., 2007. Self-Selection Models in Corporate Finance. Handbook of Corporate Finance: Empirical Corporate Finance, 47-62. Maddala, G. S., 1991. A perspective on the use of limited-dependent and qualitative variables models in accounting research. The Accounting Review. 66, 788-807.Millman, G. J., 1997. First, pop the hood. U.S. News & World Report, 122(4), 70. Retrieved from =ip,shib&db= bah&AN=9701283285&site=ehost-live.Patro, D. K., 2006. International mutual fund flows. Rutgers University working paper.Sirri, E. R., Tufano, P., 1998. Costly Search and Mutual Fund Flows. Journal of Finance. 53,?1589–1622.Wermers, R., 1999. Mutual fund trading and the impact on stock prices. Journal of Finance. 54, 581–562.Wermers, R., 2004. Is money really smart? New evidence on the relation between mutual fund flows, manager behavior, and performance Persistence. Working Paper, University of Maryland. Wilcox, R. T., 2003. Bargain Hunting or Star Gazing? Investors’ Preferences for Stock Mutual Funds. Journal of Business. 76, 645-663.Wurgler, J., 2011. On the economic consequences of index-linked investing, in W. T. Allen, R. Khurana, J. Lorsch, and G. Rosenfeld, eds., Challenges to Business in the Twenty-First Century: The Way Forward (Cambridge: American Academy of Arts and Sciences). Appendix A: Types of fund names subject to Rule 35d-1The main categories of name changes subject to Rule 35d-1 are:Names indicating an investment in a certain type of asset class (e.g. the ABC Stock Fund has to invest at least 80% of its holdings in stocks).Names suggesting investment in certain industries (e.g. the ABC Energy Fund has to invest at least 80% of its portfolio in the energy sector)Names indicating an investment emphasis in certain countries or geographic regions. This includes the mention of a specific country (e.g. Japan) or region (e.g. Latin America), as well as the inclusion of the word ‘foreign’ in the name.Names indicating that an investment company’s distribution is exempt from income tax.Names suggesting that an investment company or its shares are guaranteed or approved by the United States government. The prohibited types of names include those that use the words ‘guaranteed’ or ‘insured’ or similar terms in conjunction with the words ‘United States’ or ‘U.S. government’.Names that suggest an investment focus on any other particular type of asset. For example, the terms ‘small, mid, or large capitalization’ and ‘index’ suggest a focus on a particular type of investment, and investment companies that use these terms will be subject to the 80% investment requirement of the rule.Appendix B: Criteria for superficial and non-superficial name changesAppendix B lists the criteria used to differentiate between fundamental and superficial name changes. The former refers to a name change that is the result of (or occurs alongside) some underlying change in the fundamentals of the fund which could impact the fund’s future earnings potential. On the other hand, non-fundamental name changes are merely superficial in nature and not followed by any change in the fund, its portfolio or earnings potential. More specifically, a name change is identified as superficial in our sample if:The name change is not driven by a merger with another fund.The change does not arise because the fund is acquired by another fund.The fund manager does not change.The official investment objective (self-reported in fund prospectus) of the fund does not change.The investment focus of the fund does not change according to the Lipper classification (based on actual holdings).This results in a total of 3,680 superficial name changes and 2,315 fundamental (non-superficial) name changes in our sample. Appendix C: List of VariablesVariableNameDescriptionTNATotal Net Assets (TNA)Total Net Assets under management of fund (in $ millions).FlowNet fund flowsNet flows to a fund, calculated as in equation (1). Expressed as a percentage of the total net assets of the fund.RtnFund returnTotal fund returns over the risk-free rate.ExpManagement expenseAnnual charge to shareholders for the portfolio management services of the fund manager. It is the cost of active management and is expressed as a percentage of the total net assets of the fund.12b12b-1 feesAnnual expense for fund’s marketing & distribution costs, expressed as a percentage of the total net assets.AlpFour-factor alphaThe Carhart four-factor alpha (Carhart, 1997) is calculated using quarterly returns for the previous 12 quarters for every fund.MeanRtnMean quarterly return in past 4 quartersThe average of the quarterly total fund returns over the risk-free rate for the last 4 quarters. MeanFlowMean quarterly flow in past 4 quartersThe average of the quarterly net flows for the last 4 quarters. AgeFund ageThe number of quarters since the fund came into formation.SuperficialSuperficial dummyDummy variable that equals 1 if the name change is not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings (see Appendix B for more details).MisleadingMisleading dummyDummy variable that equals 1 if the name change is not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name.HotStyleHot style dummyDummy variable that equals 1 if the name change is a move towards (away from) a name reflecting a particular investment style when the associated style premium is positive (negative) over the last two quarters.LoopholeLoophole dummyDummy variable that equals 1 if the name change exploits a loophole in the new regulation (i.e., if the new name includes the terms ‘value’, ‘growth’, ‘balanced’, ‘international’ or ‘global’).Appendix D: Performance comparison of different fund categoriesThis table reports mean characteristics and differences on two measures of performance for a number of fund categories. Performance is measured by total fund returns over the risk-free rate and the four-factor alpha generated by funds in the previous quarter. Data are aligned on the date of the name change. Performance is measured for four fund categories: all funds, all name-change funds, funds that undergo only superficial name changes and funds that undergo only misleading name changes. Superficial name changes refer to changes not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings. Misleading name changes refer to changes not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name. The p-values are based on the t-test on the differences and are reported in parentheses. ***, ** and * refer to significance at the 1%, 5% and 10% levels respectively.Raw returnsFour-factor alpha(1) All funds1.15%-0.68%(2) Name change funds0.27%-1.12% (2) – (1) Difference-0.89%***(0.00)-0.44%**(0.03)(3) Superficial name change funds0.43%-1.33% (3) – (1) Difference-0.73%***(0.01)-0.65%***(0.01)(4) Misleading name change funds0.68%-0.93% (4) – (1) Difference-0.48%(0.36)-0.26%(0.60)Figure 1: Frequency of Name Changes over Sample PeriodThe number of name changes per year is plotted for the entire sample period, starting from Q4-2002 to Q4-2011 inclusive. The dotted bars represent the total number of name changes. The striped bars show the number of superficial name changes (i.e., changes not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings) while the solid bars depict the number of misleading name changes (i.e., changes not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name).Table 1: Descriptive Statistics of Name-change FundsPanel A reports mean and median characteristics for name-change funds and their counterparts across our sample period (Q4/2002 to Q4/2011). Figures are based on fund averages for funds that make multiple name changes. Data are aligned on the date of the name change. Hence lagged values represent fund characteristics in the quarter before the fund changed its name (unless otherwise specified). Lagged fund total net assets (TNA) refer to the assets under management in the quarter before the name change. Fund flow is defined similar to Sirri and Tufano (1998) as {[TNAt – TNAt-1(1 + rt)] - MGTNAt} / TNAt-1. Fund lagged return refers to the total fund returns over the risk-free rate (i.e., total excess return) in the prior quarter. The management expense is the ratio of the total investment that shareholders pay to the fund manager for their portfolio management expertise. The 12b-1 fee is an expense that shareholders pay every year to cover a fund’s marketing and distribution costs. The value reported is the annual percentage of the fund’s total net assets that is attributed to these costs. The Carhart (1997) four-factor alpha is calculated using quarterly returns for the 12 quarters prior to the name change. Mean quarterly return is the average of the quarterly returns for the last four quarters. The mean fund flow is an average of the quarterly flow for the last four quarters. The fund age is the number of quarters a fund has been in existence up to the name change event. Panel B reports the number of name-change funds in each quintile based on breakpoints derived from the characteristics of all non-change funds. For each name change, the quintile breakpoints are determined relative to all other funds on the event date. Quintile 1 is the smallest and 5 the largest. The p-values are also reported, obtained from a χ2 one-sample test for uniform distribution of name-change fund counts across the five quintiles.Panel A: Descriptive statistics on fund characteristicsName-Change FundsOther FundsMeanMedianMeanMedianLagged fund total net assets ($millions)457.38122.951143.79178.8Lagged net fund flow (%)1.05-1.101.64-0.62Lagged fund return (%)0.271.791.172.32Lagged management expense (% per year)0.930.980.941.09Lagged 12b-1 fees (% per year)0.300.250.190.08Lagged four-factor alpha (%)-1.12-0.40-0.67-0.38Mean quarterly return in past 4 quarters (%)1.282.111.221.97Mean quarterly flow in past 4 quarters (%)1.40-0.522.04-0.15Fund age (in quarters)34.9228.6739.3631.00Panel B: Distribution of characteristics of name-change fundsQ1Q2Q3Q4Q5p-valueLagged fund total net assets ($millions)5686085604422880.00Lagged net fund flow (%)6134703994384570.00Lagged fund return (%)632520503518 5020.00Lagged Management Expense (% per year)6383654895396320.00Lagged 12b-1 fees (% per year)4181064235578960.00Lagged four-factor alpha (%)5244905185203980.00Mean quarterly in past 4 quarters (%)5044954735344720.00Mean quarterly flow in past 4 quarters (%)6224864304184660.00Fund age (in quarters)5765625935294170.00Table 2: Determinants of Mutual Fund Name ChangesThis table employs a logistic regression to examine the relationships between fund characteristics and the likelihood of a fund changing its name. Panel A investigates what types of funds choose to change names. It reports the results from a cross-sectional logistic regression where the dependent variable is a dummy that is assigned a value of 1 for a name change event and 0 for all other fund-quarters (for first column of results). The dummy variable takes a value of 1 for different categories of name changes (superficial, non-superficial, misleading and non-misleading) for the other columns. Superficial name changes refer to changes not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings, whereas non-superficial refer to all remaining changes. Misleading name changes refer to changes not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name, whereas non-misleading refers to all remaining changes. The independent variables in all cases remain the same and are as defined in Table 1. Panel B investigates when funds choose to change their names. In this case, a time-series logistic regression is run only for the sample of funds that change names. The dependent variable is a dummy that take the value of 1 at the name-change quarter and 0 at all other quarters (and similarly for different categories of name changes). Results where the dummy is 1 for different categories of name changes are also reported. Explanatory variables are as described earlier. Values reported are the marginal effects at the mean and corresponding p-values are based on the Z-test (reported in parentheses). Statistical significance at the 1%, 5% and 10% level is denoted by ***, ** and * respectively.Panel A: Cross-sectional Logistic RegressionAllSuperficialNon-superficialMisleadingNon-misleadingLog of lagged total net assets ($ millions)0.000(0.94)-0.001(0.31)0.000(0.35)0.000(0.75)0.000(0.97)Lagged net fund flow (% of TNA)-0.585(0.56)-4.790(0.52)-0.114(0.78)-0.477*(0.08)0.027(0.97)Lagged fund returns (%)-0.483(0.89)0.621(0.83)0.004(0.99)0.545(0.56)-1.112(0.73)Lagged management expense (% per year)-0.768***(0.01)-0.852***(0.00)-0.067(0.44)-0.151***(0.00)-0.601(0.25)Lagged 12b-1 fees (% per year)6.728***(0.00)4.614***(0.00) 1.675***(0.00)0.886***(0.00)5.457***(0.00)Mean quarterly return over past 4 quarters (%)0.908(0.88)2.298(0.57)-0.997(0.73)0.176(0.90)1.333(0.79)Lagged four-factor alpha (%)-5.627**(0.05)-2.923(0.21)-2.296*(0.06)-0.570(0.16)-4.94*(0.07)Mean quarterly fund flow over past 4 quarters (%)-1.966(0.13)-0.754(0.40)-1.356**(0.02)0.013(0.96)-2.224**(0.05)Log of fund age (age in quarters)-0.006**(0.00)-0.003**(0.04)-0.181*(0.07)0.000(0.39)-0.521***(0.00)Pseudo R211.04%10.90%12.07%11.55%10.98%No. of Observations71,57171,57171,57171,57171,571Table 2 (Continued)Panel B: Time-series Logistic Regression for Name-Change FundsAllSuperficialNon-superficialMisleadingNon-misleadingLog of lagged TNA ($ millions)-0.003***(0.00)-0.003***(0.00)-0.001*(0.08)-0.005***(0.00)-0.003***(0.00)Lagged fund net flow (%)-0.779(0.41)-0.483(0.68)-2.345(0.23)-5.144*(0.08)-0.161(0.88)Lagged fund returns (%)-7.092***(0.00)-2.821(0.32)-14.659***(0.00)1.655(0.86)-8.297***(0.00)Mean quarterly return over past 4 quarters (%)4.475(0.25)0.645(0.90)11.653*(0.08)9.308(0.55)4.325(0.30)Lagged four-factor alpha-0.244(0.81)-2.287(0.13)2.036(0.20)-0.499(0.92)-0.332(0.76)Mean quarterly flow over past 4 quarters (%)-1.390(0.26)-1.121(0.47)-1.937(0.38)2.365(0.57)-2.005(0.13)Log of Fund Age-0.007***(0.00)-0.008***(0.00)-0.007***(0.00)-0.001(0.83)-0.009***(0.00)Time dummiesYesYesYesYesYesPseudo R24.93%5.02%7.48%3.35%5.02%No. of Observations38,79726,69311,1746,66831,530Table 3: Abnormal Fund flows after Name ChangesThis table reports the abnormal flows following name changes. Panel A reports the results of the counterfactual construction following from the endogenous switching regression analysis, as outlined in section 4.3. This corresponds to equation (5). The abnormal flows are calculated for name change funds by subtracting the computed counterfactual flows from the observed flows. Counterfactual flows refer to the constructed flows to name-change funds under the (unobserved) alternative scenario if these funds had not changed names. The analysis is undertaken for normalized flows for the quarter after the name change as well as for the mean quarterly flow in the year after the event. Panel B presents the difference across the two categories of interest: (i) misleading and non-misleading changes (ii) superficial and non-superficial flows for both measures of fund flows. Superficial name changes refer to changes not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings, whereas non-superficial refer to all remaining changes. Misleading name changes refer to changes not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name, whereas non-misleading refers to all remaining changes. The p-values are based on the t-test on the differences. ***, ** and * refer to significance at the 1%, 5% and 10% levels respectively. Panel A: Counterfactual analysis for name-change fundsMean flow over 1 quarter Mean (quarterly) flow over 1 yearActual CounterfactualDifferencep-valueActual CounterfactualDifferencep-valueAll name changes1.33%-7.78%8.89%0.001.15%-2.39%3.22%0.00Misleading0.74%-7.59%7.67%0.000.43%-2.23%2.08%0.00Non-misleading1.45%-7.88%9.09%0.001.29%-2.42%3.45%0.00Superficial1.56%-7.61%8.85%0.001.40%-2.19%3.24%0.00Non-superficial0.99%-8.23%8.86%0.000.79%-2.74%3.19%0.00Panel B: Test for difference between cumulative abnormal flowsMean abnormal flow over 1 quarter Mean abnormal (quarterly) flow over 1 yearMisleading name changes7.67%2.08%Non-misleading name changes9.09%3.45% Difference-1.41% -1.37%** p-value for difference0.110.03Superficial name changes8.85%3.24%Non-superficial name changes8.86%3.19% Difference0.00%0.06% p-value for difference0.990.91Table 4: Determinants of Abnormal Flows to Name Change FundsThis table examines the determinants of abnormal flows to funds in the subsequent 1 and 4 quarters after the name change event. Panel A reports the results where the dependent variable is the abnormal flows 1 quarter after the name change event while Panel B reports the results where the dependent variable is the mean quarterly abnormal flows over 4 quarters after the name change event. The explanatory variables are as follows: a superficial dummy that equals 1 if the name change is not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings, a misleading dummy that equals 1 if the name change is not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name, a hot style dummy that equals 1 if the name change is a move towards (away from) a name reflecting a particular investment style when the associated style premium is positive (negative) over the last two quarters, a loophole dummy that equals 1 if the name change exploits a loophole in the new regulation, the log of TNA in the quarter before the event, the quarterly fund flow (calculated as {[TNAt – TNAt-1(1 + rt)] - MGTNAt} / TNAt-1), the excess returns in the quarter before the name change, the four-factor alpha calculated over quarters -12 to -1 before the name change, the log of fund age in quarters, and the change in management expense ratio and 12b-1 fees from one quarter before to one quarter after the name change. The p-values are based on the t-test and reported in parentheses. ***, ** and * refer to significance at the 1%, 5% and 10% levels respectively. Panel A: Abnormal flows 1 quarter after name change as dependent variableModel 1Model 2Model 3Model 4Model 5Model 6Superficial dummy0.003(0.70)0.003(0.70)0.003(0.70)0.003(0.74)0.000(0.98)0.004(0.53)Misleading dummy-0.013(0.13)-0.015*(0.09)-0.010(0.27)-0.012(0.20)-0.009(0.35)Hot style dummy0.023*(0.08)0.024*(0.06)0.024*(0.06)0.024*(0.08)Loophole dummy-0.014**(0.02)-0.014**(0.04)-0.004(0.53)Log of lagged TNA-0.005*(0.06)-0.005*(0.06)-0.005*(0.06)-0.005*(0.07)-0.004(0.15)-0.004(0.12)Lagged fund flow (%)3.335(0.29)3.322(0.29)3.292(0.29)2.955(0.34)2.257(0.47)0.970(0.76)Lagged returns (%)30.823***(0.00)33.999***(0.00)Lagged 4 factor alpha (%)-11.232*(0.08)-10.622(0.16)Log of age-0.008(0.13)-0.005(0.34)Change in 12b-1 fee (%)4.525(0.19)Change in management expense (%)-7.776***(0.00)Constant0.101***(0.01)0.102***(0.01)0.097***(0.00)0.102***(0.00)0.182***(0.00)0.168***(0.00)Time dummiesYesYesYesYesYesYesAdjusted R25.82%5.82%6.02%6.23%7.47%8.74%No. of Observations1,9831,9831,9831,9831,9831,644Panel B: Mean quarterly abnormal flows over 4 quarters after name change as dependent variableModel 1Model 2Model 3Model 4Model 5Model 6Superficial dummy0.004(0.40)0.004(0.41)0.004(0.41)0.004(0.42)0.003(0.63)0.002(0.71)Misleading dummy-0.012*(0.06)-0.013**(0.05)-0.011**(0.08)-0.012**(0.06)-0.009(0.14)Hot style dummy0.004(0.58)0.004(0.57)0.005(0.54)0.006(0.46)Loophole dummy-0.003(0.49)-0.003(0.59)-0.001(0.86)Log of lagged TNA-0.003**(0.05)-0.003**(0.05)-0.003**(0.05)-0.003**(0.05)-0.002(0.26)-0.002(0.18)Lagged fund flow (%)-1.179(0.57)-1.200(0.56)-1.202(0.56)-1.270(0.54)-2.175(0.30)-3.097(0.15)Lagged returns (%)18.640***(0.00)23.583***(0.00)Lagged 4 factor alpha (%)-7.040(0.12)-6.615(0.21)Log of age-0.010***(0.01)-0.009***(0.01)Change in 12b-1 fee (%)0.771(0.77)Change in management expense (%)-4.387***(0.00)Constant0.057***(0.00)0.057***(0.00)0.057***(0.00)0.058***(0.00)0.124***(0.00)0.129***(0.00)Time dummiesYesYesYesYesYesYesAdjusted R22.27%2.46%2.47%2.50%3.68%5.87%No. of Observations2,0182,0182,0182,0182,0181,667Table 5: Difference in average Fund Characteristics around the Name ChangeThis table reports differences in the average quarterly fund characteristics in the year before the name change event to the year after. Quarterly excess returns refer to the total fund returns over the risk-free rate for the quarter. The four-factor alpha is calculated using quarterly returns for the previous 12 quarters. The management expense is the portion of the total net assets that shareholders pay for the fund’s management services (for the year). The 12b-1 fee is an expense that shareholders pay every year to cover a fund’s marketing and distribution costs. The value reported is the annual percentage of the fund’s total net assets that is attributed to these costs. Results are also reported separately for superficial and misleading name changes. Superficial name changes refer to changes not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings. . Misleading name changes refer to changes not accompanied by a corresponding change in portfolio holdings to reflect the investment style suggested by the new name. The p-values reported are based on the t-test on the differences. AllSuperficialMisleadingDifferencep-valueDifferencep-valueDifferencep-valueQuarterly excess returns (%)0.020.90-0.440.05-1.380.00Four-factor alpha (%)1.140.130.860.33-0.170.57Management expense (% per year)0.010.180.030.000.000.4712b-1 fees (% per year)0.010.000.000.01-0.020.01Table 6: Additional Mutual Fund Name ChangesThis table reports results on the 567 additional name changes in our sample. A name change is classified as additional if it is not the first name change undertaken by the fund in our sample period. We repeat the analysis in Table 3 to calculate abnormal flows towards additional name changes. We also distinguish between superficial and non-superficial additional name changes and calculate differences. Superficial name changes refer to changes not accompanied by a change in the fund’s ownership, management, investment strategy or portfolio holdings, whereas non-superficial refer to all remaining changes. Results are presented for both normalized flows for the quarter after the name change as well as for the mean quarterly flow in the year after the event. The p-values reported are based on the t-test on the differences. Mean flow over 1 quarter Mean (quarterly) flow over 1 yearActual CounterfactualAbnormalp-valueActual CounterfactualAbnormalp-value(1) All additional changes-0.75%-8.28%6.79%0.00-0.43%-2.74% 1.84%0.00(2) Non-additional changes1.99%-7.65%9.59%0.001.65%-2.25%3.71%0.00(1) – (2) difference-2.79%(0.00)-1.87%(0.00)(3) Additional superficial-0.75%-8.13%6.47%0.00-0.22%-2.62%1.90%0.00(4) Non-additional superficial 2.32%-7.42%9.66%0.001.93%-2.04%3.70%0.00(3) – (4) difference-3.20%(0.00)-1.80%(0.01)(5) Additional Non-superficial-0.74%-8.56%7.35%0.00-0.77%-2.96%1.73%0.01(3) – (5) difference-0.88%(0.56)0.17%(0.86) ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download