I investigate mutual fund flows at the individual fund and at the fund family level. At the individual, I use SEC filings to decompose fund flows into inflows and outflows. This decomposition of net flows into its component parts provides a way to examine differences in how search costs and investor learning affect investors who are entering a fund (or adding to their investments) versus those investors who are leaving a fund (or decreasing their investments). I then examine the effect of the existence of an advertisement for the fund on these investors. At the mutual fund family level, I examine how the characteristics and performance of mutual fund families affect the flows to the family as a whole. I then examine the effects of advertising expenditures on flows to the fund family.

The importance of mutual funds for individual investors has increased in recent decades. This becomes apparent when looking at the increased share of households owning mutual funds. These mutual fund investors usually want to receive a return which is above or at least close to the mutual fund’s benchmark. Consequently, investors want to invest in those funds which will show these patterns in the future. Some of these mutual funds receive much attention, since they generate extraordinary high performance. But the question that remains is whether it is possible to predict such performance before funds exhibit such outstanding performance. In the past, mutual fund investors focused extensively on performance or performance linked patterns, like the Morningstar star rating, and thus chased past performance. This seems surprising since performance persists only over a short time and is more persistent to weak mutual funds (1 and 2 star rated) than well performing mutual funds. Thus, chasing past performances seems to be a rather inferior strategy. Therefore, investors should try to identify alternative tools showing a high correlation to future mutual fund performance. In this book, mutual funds are analysed, especially open-end mutual funds and actively managed mutual funds. The main focus is on what purpose and usefulness active investments have and whether performance is persistent and what the determinants of mutual fund flows are. Moreover, some alternative measures will be introduced by explaining which attributes or methods should be used and avoided when selecting mutual funds.

How to prevent runs on open-end mutual funds? In recent years, markets have observed an innovation that changed the way open-end funds are priced. Alternative pricing rules (known as swing pricing) adjust funds’ net asset values to pass on funds’ trading costs to transacting shareholders. Using unique data on investor transactions in U.K. corporate bond funds, we show that swing pricing eliminates the first-mover advantage arising from the traditional pricing rule and significantly reduces redemptions during stress periods. The positive impact of alternative pricing rules on fund flows reverses in calm periods when costs associated with higher tracking error dominate the pricing effect.

This dissertation investigates the determinants of mutual fund flows and mutual fund performance. The first chapter examines the response of fund investors to style volatility and the impact of style volatility on the flow-performance relationship. Three main empirical findings are obtained using both a portfolio approach and a multivariate regression approach. First, I find that there is a significant positive relationship between the style volatility and the subsequent fund flows to mutual funds. This finding can be interpreted as either fund managers having style timing ability or fund managers catering to investors preferences or tastes. Second, the positive relationship between past style volatility and fund flows is less pronounced for funds with superior past performance. Lastly, fund style volatility has a dampening effect on the flow-performance relationship: the flow-performance sensitivity weakens by 12% when the past style volatility increases by one standard deviation. It is likely that performance is perceived as a less informative signal of investment ability for fund managers who follow inconsistent styles over time. The second chapter studies how the response of fund investors to past risk varies over business cycles. I employ the NBER boom indicator, the Consumer Sentiment Index, and the National Activity Index to proxy for economic conditions. I find that mutual fund investors react differently to risk across economic environments. Funds with more volatile past returns discourage fund investors. The investors’ demand for actively managed funds is higher under good market conditions. Fund flows are less responsive to risk during expansionary economic periods. This finding may indicate that fund investors are risk averse and become less risk averse in good market states. The third chapter empirically examines whether mutual fund performance is affected by prior family performance. I propose two testable hypotheses: the information and resource sharing hypothesis and the cross-fund subsidization hypothesis. The empirical findings suggest that there is a significant positive relationship between prior family performance and subsequent fund performance. This finding is consistent with the hypothesis that mutual funds in the same family share informational resources. This positive relation also justifies the finding in the mutual fund flow literature that fund flows are higher for funds with higher past family performance. Furthermore, I find that the predictive power of the prior family performance is stronger in larger fund families.

This thesis uses the recent surge in ETFs as an opportunity to explore the determinants of ETF flows as opposed to mutual fund flows. Using monthly observations between 2009 and 2015 of 1'413 mutual funds and 328 ETFs of US equity funds, an autoregressive process and piecewise linear regression was applied. The results confirm several previous findings regarding the high degree of flow persistence recorded in mutual funds and suggest a similar, but slightly weaker degree of persistency in ETF flows. Further, ETF flows show a significantly higher sensitivity towards management fees and volatility in fund performance. Lastly, the convex flow-performance relation documented in mutual funds was confirmed using the data at hand, and - surprisingly - also found in ETF flows.