Investors routinely pour cash into recently launched funds. Is that a mistake?
By Lee Davidson, Madison Sargis, and Tim Strauts
The following article was part of a larger research study on the subject. To get a copy of the full white paper, email the authors at email@example.com.
It is easy to dismiss new funds as an immature, unproven, and unimportant segment of the asset-management industry. But the truth of the matter is that the asset-management industry is dominated by new funds. Globally, we find that new funds account for the preponderance of new asset flows. In 2015, global fund flows reached approximately $516.4 billion across the three asset classes in our study—equity, fixed income, and allocation. Of the total, new funds with less than 12 months of track record accounted for $379 billion, or 73% of all flows. Even in years when the industry experiences net outflows, new funds continue to garner assets. In 2014, new funds grabbed $316 billion in net inflows compared with negative $526 billion in net outflows for funds with greater than 12 months of track record.
We hope, therefore, that it does not come as a surprise why the patterns of flows into new funds should be of interest to us as researchers. It is simply too big and too important to ignore. Furthermore, we are particularly concerned about investor behavior when investing in these unproven products. Do we find that investors prefer what’s good for them?
In this paper, we explore the relationship between observed investor preferences for and eventual investor outcomes in newly launched funds. To conduct this exploration, we built two models that unpack the historical correlations between both forward risk-adjusted returns and forward cumulative fund flows with observed fund characteristics, economic regimes, and category environs. In isolation, each model offers us a multivariate view of the factors that drive future flows and risk-adjusted returns for newly launched funds. By comparing the outputs from the two models, we can also identify potential conflicts of interest for asset managers. A variable that has tended to result in higher flows but lower returns, for example, socially responsible investing, presents a difficult choice for asset managers over whose interests to prioritize.
This tension between what investors have preferred and the outcomes investors are seeking also highlights poor choices on the part of the investing populace. We find that investors have been inconsistent with how they allocate their assets among newly launched funds. For example, we find that investors prefer cheaper funds that, not surprisingly, tend to result in better outcomes. But we also find that investors prefer funds that invest in more-popular, familiar stocks that have done well recently when the opposite choice would have resulted in better risk-adjusted returns.
Newly launched funds are of particular interest to us as they often lack information that we know investors use to make decisions. From prior work, we have observed that investors exhibit strong preferences for funds with strong performance track records. While this is perhaps not surprising, it does raise the question of what information investors use to make an investment decision when a performance track record is unavailable. Despite the lack of information on newly launched funds, we have already presented evidence that new funds can garner immense assets almost immediately.
Given the lack of relevant data early on in the life cycle of a fund, variable selection and variable definition were critical choices in our study. Variables included in the model had to be both widely available in our sample and hypothesized to be relevant drivers. We included 23 variables in the fixed-income and allocation models. In the equity model, we were able to expand this number to 39 because of the quality of Morningstar’s equity data and portfolio holdings database. Many variables included in the model are to be expected: fees, firm-level characteristics, index fund, socially responsible, and portfolio disclosure. However, some variables are less common. Category-level data points such as market concentration and firm market share by category were constructed specifically to examine the role of category structure on newly launched fund outcomes. Furthermore, we also expanded the use of our ownership data by building data points to examine whether a fund manager behaves in a similar fashion to other types of managers. An example of this is the Manager CFA Ownership data point, which captures the extent to which the manager of a newly launched fund is buying stocks in a manner similar to the typical manager who is a CFA charterholder. In general, we attempted to capture various perspectives of a new fund by focusing on data categories such as fund structure, manager demographics, firm reputation, competition within category, portfolio disclosure, portfolio style, and economic regimes.
For the purposes of this paper, we define newly launched funds as any fund with less than 12 months of track record. We do not consider a newly offered share class to be a new fund. We are only interested in brand-new funds. We include open-end funds and exchange-traded funds in our sample. All explanatory data that we use would have been available within that first year. The variables we are trying to explain are forward 36-month risk-adjusted returns and forward 36-month cumulative fund flows. The study is global and begins in January 2005. The last new funds to be included in our study were launched by March 2013.
Findings on New Funds
Managers who own their funds have historically performed better and garnered more assets. Fund management ownership is an excellent signal for investors. Data from our study indicates that portfolio managers who have their financial interests aligned with investors run higher-quality funds. Investor preference for such ownership outpaces the benefit of higher risk-adjusted returns. A typical equity fund will have on average 5.5% higher risk-adjusted returns and will move up 6.5% percentiles for flows within its category. For fixed income and allocation, the benefit is more dramatic, 1.9% and 4.0% increases in risk-adjusted returns, with a 6.9% and 8.9% percentile flows increase, respectively.
We did not differentiate between the levels of manager ownership, only noting whether or not a single manager had at least $1 in the fund. We noticed a significant number of funds report no manager investment in their funds. By doing so, the fund is effectively sending a signal that the management team’s financial interests are not aligned with their end investors’ and they are not willing to do so, even at a minimal cost. Not doing so suggests that the funds may be losing out on future business.
CFA Difference: Charterholders achieve better outcomes and are preferred by investors. Investor preferences for CFA-designated portfolio managers follows similarly to their preferences for manager ownership. While the average increase in fund returns across asset classes ranges from 1.1% to 1.5%, the expected average flows varies among fund type. Fixed-income investors display the strongest preference for managers who are CFA charterholders as the average benefit is a 9.8% increase in category flow percentiles. Equity and allocation funds on average receive a 4.7% and 2.5% percentile increase, respectively.
Investors have imperfect knowledge about a manager’s capabilities, so they are likely using the CFA charter as a proxy for skill and education. Our study shows the signal is a beneficial metric to investors from the perspective of improving their long-term performance, too. Interestingly, the preference varies dramatically across asset classes. This tells us the importance of a more educated manager differs for investors depending on the investment type. Yet, the effects in terms of outperformance are very similar, regardless of asset class. Regardless of the magnitude, having a fund managed by a CFA charterholder has resulted in higher flows and higher star ratings. The interests of the asset manager correspond to the investor.
Shifting Gender Preferences: Female portfolio managers tend to garner more assets.
The inherent trait of a manager’s gender affects flows into a fund. To determine an investor’s perception of a portfolio manager’s gender, we calculated the probability of being female, given a first name and birth year. For a team-managed fund, we used the highest probability of all the fund’s managers. Instead of using a binary indicator variable, the continuous probabilities can tell us about an investor’s perception of gender. If investors can automatically assume a portfolio is managed by a woman, do they prefer that fund more than a fund with less clear information regarding the manager’s name? In doing so, we can test whether there is an overall investor bias for female portfolio managers.
Our study suggests gender does matter to investors. Flows follow female portfolio managers for equity and fixed-income asset classes. Our hypothesis for why there is this preference for female management corroborates previous Morningstar studies with our data. Last year, the “Fund Managers by Gender” report determined that women are underrepresented in the fund management industry relative to other professions (Lutton, Davis). Therefore, we assume women face significant headwinds in advancing their careers in the financial industry. So, the average woman who does advance to become a portfolio manager should be higher-performing than the average male portfolio manager. Our reasoning implies a female portfolio manager signals to an investor higher management skill, which is represented by a positive association between flows and gender. As the demographics of who is in control of private wealth change to become more split between men and women, we expect to see preference for female portfolio managers to be positive and grow over time (Lutton, Davis).
While we can logically reason why an investor may prefer a female portfolio manager, we find mixed effects when studying gender’s correlation to higher forward star ratings among new funds. We do not expect gender to be a proxy for skill.
High fees hurt flows and future risk-adjusted returns. Fees are one of the most consistent drivers of flows and returns. Higher fees led to lower returns and lower flows across all asset classes for new funds; however, the magnitude relative to other factors is smaller than expected. In our study we control for manager traits, fund structure, portfolio style, firm and category characteristics and macroeconomic environments.
Unsurprisingly, fees play an important role in the forward star ratings. A category’s most expensive fund could lose out on 3% in terms of risk-adjusted returns for equity and allocation funds compared with the cheapest fund. The effect is less dramatic for fixed-income funds, as the difference is only 0.57% risk-adjusted returns on average. The results of the study line up with recent Morningstar research that outlines how fees have a predictive power of success. Therefore, knowing fees could be used as a proxy for forward success, we were surprised to see fees were only, at minimum, the ninth most economically meaningful driver of flows across asset classes. Lowering a fund’s expense ratio one percentile relative to the category will boost a fund’s forward cumulative flows only a fraction of a percentile for all asset classes.
Morningstar coverage is suggestive of higher flows and better investor outcomes.
Morningstar analysts are skilled at finding successful new funds, and investors take notice. Across asset classes, funds covered by Morningstar generate higher risk-adjusted returns. The largest effect is in allocation, where thre is a maximum increase of 4.7%. The smallest effect is in equity, where we see only a 1.7% maximum benefit.
Investors look to Morningstar to perform due diligence and then often follow our advice. The effect of a Morningstar Analyst Rating is immense. The sooner the analyst starts rating the fund, the more time investors have to digest the information and take action. If Morningstar rates an equity fund in the first month, the fund moves into a 14.4% higher category flow percentile. The effect is equally large for allocation funds, where the percentile increase is 13.5%. Interestingly, investors are not valuing the rating as much for a fixed-income fund. During the same time period, the percentile increase was only 0.4%.
Style tilts that have resulted in good investor outcomes are not widely preferred by investors. This is the most disappointing finding in our paper from two perspectives: 1) investors don’t seem to make good choices when it comes to choosing new funds, and 2) bad choices create a conflict of interest for asset managers. Broadly speaking, we find that investors in new funds prefer to give their money to funds that invest in popular companies that have done well recently even though almost all evidence suggests that the opposite choices result in better investor outcomes. We were able to study this only for equity funds, but, given the magnitude and persistence of these effects in our equity fund sample, we don’t hold much optimism for reversed results in fixed-income and allocation asset classes.
Second, the good news for investors is that their preferences have generally paid off in better outcomes. Funds that exhibit the types of traits listed above are generally the better cohorts of funds from the investor’s perspective. The main exception to this comes from investor preference for style tilts, which is directly contrary to their best interests. If portfolios have been disclosed, the investing populace tends to place a premium on funds that buy popular, large-cap, overvalued, and liquid stocks that have done well recently. Investors appear to ask themselves, “Have I heard of these stocks?” and respond with additional flows to the new fund when the answer is “yes.” Unfortunately, in almost all instances, we would expect the opposite choice to result in a higher expected return.
We conclude this paper by highlighting our potential future work. As with most papers, several questions were raised by conducting this analysis. Two stand out:
We anticipate studying these questions and more in the coming months.