Business

ETF Flows And Underlying Stock Returns: The True Cost Of NAV-Based Trading

ETF Flows And Underlying Stock Returns: The True Cost Of NAV-Based Trading via SSRN 

Friedrich Osterhoff

Technische Universität München (TUM) – Department of Financial Management and Capital Markets

Maximilian Overkott

Technische Universität München (TUM) – Department of Financial Management and Capital Markets

February 22, 2016

Abstract:

In this paper, we examine whether the creation or redemption of exchange-traded fund (ETF) shares has a measurable and significant effect on the underlying stocks’ returns in the closing auction. Our findings show that ETF flow-related stock transactions significantly affect stock prices. We provide empirical evidence showing that creations-redemptions of ETFs that replicate indices in the Deutsche Börse AG German Stock Index (DAX) family have a highly significant and economically viable positive/negative effect on abnormal returns of underlying stocks in the closing auction. The effect is particularly pronounced in small stocks and on trading days that are, generally, bullish. We argue that the Authorised Participant has a motivation exploit this inefficiency, and that, given the sizable additional annual earnings of up to EUR 53000 per stock, active price manipulation during the closing auction should be an attractive option. Hence, dealing in ETF shares on the primary market, for example through net asset value-based orders, might entail hidden costs for investors that, until now, have not been recognised in the literature, or, perhaps, not even by investment professionals.

ETF Flows And Underlying Stock Returns: The True Cost Of NAV-Based Trading – Introduction

Exchange-traded funds (ETFs) are widely acknowledged to be one of the most successful financial innovations of the past few decades, a perception that is supported by the substantial growth of assets under management (AuM) over the past several years, amounting to USD 2.66 trillion globally as of August 2015 (Deutsche Bank Research, 2015), and the growing share of ETF-related transactions in total trading volume. With ETFs playing such an increasingly vital role in capital markets, a thorough understanding of how they might affect underlying securities, and, in that way also markets as a whole, is of significant interest to both investment professionals and financial researchers.

The effect that ETFs have on the quality of underlying securities is not entirely new to theoretical or empirical financial research. In that field, existing literature on ETFs often focusses on a fund’s inception and its effect on stock liquidity (e.g. Hedge and McDermott (2004), Yu (2005), Richie and Madura (2007), and De Winne et al. (2013)), volatility (e.g. Lin and Chiang (2005) and Ben-David et al. (2014)), or long-term valuation (e.g. Madura and Ngo (2008) and Bae et al. (2013)). Other studies also confirm the effect of fund flows on stock prices (e.g. Edelen and Warner (2001), Yu (2005), Coval and Stafford (2007), and Jotikasthira et al. (2012)). A growing yet still relatively small fraction of studies acknowledge the structural differences between ETFs and mutual funds, the in-kind creation-redemption of ETF shares in particular; thus, there is a need for research that clearly and specifically focuses on ETFs in order to more fully understand how their fund flows might affect underlying securities (e.g. Kalayc?o?lu (2004), Staer (2014)).

ETF Flows

This present study of German ETFs and their relation with and impact on their underlying stocks aims to extend the research that recognises ETFs not as a sub-class of mutual funds but as a separate class of assets. We also seek to shed new light on the relationship between ETF flow-related trading volume and stock returns. In contrast to previous studies that describe the effect of ETFs on their respective underlying stocks at an aggregated index- or market-level, we emphasise the need to identify the different levels of influence that impact the size and liquidity of a stock. Toward that end, we investigate ETF flow effects on returns at the most granular level possible, namely that of individual stocks. The variable in our approach is the abnormal stock return in the closing auction in the German stock market. By using a controlled environment, such as the closing auction, it is less likely that general market movements or news would drive abnormal returns in our sample.

Our findings show that ETF flow-related stock transactions significantly affect stock prices. We provide empirical evidence showing that creations-redemptions of ETFs replicating indices in the Deutsche Börse AG German Stock Index (DAX) family have a highly significant and economically viable positive/negative effect on abnormal returns of underlying stocks in the closing auction. The effect is particularly pronounced in small stocks and on trading days that are, generally, bullish. We argue that the Authorised Participant (AP), might be able to exploit this inefficiency, and that, given the sizable additional earnings (according to our models), active price manipulation might be an attractive option. Dealing ETF shares, for example through net asset value (NAV)-based orders, might entail hidden costs that, until now, have not been recognised in the literature, or, perhaps, not even by investment professionals.

ETF Flows

The fact that these effects are measurable and robust for one of the world’s largest and presumably most efficient stock markets increases concerns about the even greater impact they could have in less-developed markets and the hidden costs for investors, which could be sizeable.

The remainder of this paper is structured as follows. Section 2 provides an overview of the relevant literature with a focus on the correlation between fund flows and stock prices. Section 3 presents the empirical components of the study. Section 3.1 provides information on the institutional background. Section 3.2 describes the data and research design, Section 3.3 presents the results and discusses the implications of our findings and the robustness of our models. In Section 4, we provide our concluding remarks.

ETF Flows

See full PDF on SSRN