Exchange-traded Funds Returns Research Articles

Forecasting ETF Performance: A Comparative Study of Deep Learning Models and the Fama-French Three-Factor Model

The global financial landscape has witnessed a significant shift towards Exchange-Traded Funds (ETFs), with their market capitalization surpassing USD 10 trillion in 2023, due to advantages such as low management fees, high liquidity, and broad market exposure. As ETFs become increasingly central to investment strategies, accurately forecasting their performance has become crucial. This study addresses this need by comparing the efficacy of deep learning models against the traditional Fama-French three-factor model in predicting daily ETF returns. The methodology employs eight artificial neural network architectures, including ANN, LSTM, GRU, CNN, and their variants, implemented in Python and applied to data ranging from 2010 to 2020, while also exploring the impact of additional factors on forecast accuracy. Empirical results reveal that LSTM and the Fama-French three-factor model exhibit a superior performance in ETF return prediction. This study contributes to the literature on financial forecasting and offers practical insights into investment decision making. By leveraging advanced artificial intelligence techniques, this study aims to enhance the toolkit available for ETF performance analysis, potentially improving investment strategies in this dynamic market segment.

Delving into the Exchange-Traded Funds (ETFs) Market: Understanding Market Efficiency

Exchange-traded funds (ETFs) are the most popular products in the financial sector today. There is extensive literature on the multifractal analysis of some stock markets, but not about the multifractal behaviour of the ETF market. This study examines the efficiency of stock index ETFs worldwide from an Efficient Market Hypothesis (EMH) perspective, using the ETFs: Ishares Msci World ETF (URTH), Ishares Russell 1000 ETF (IWB), SPDR S&P 500 ETF TRUST (SPY), Ishares Global Clean En. ETF (ICLN), Ishares USD Green Bond ETF (BGRN), from 1 January 2021 to 24 May 2024. It analyses a pre-conflict and a geopolitical conflict to uncover distinct patterns of behaviour reflecting significant changes in market conditions. Before the conflict, the Ishares MSCI World, Ishares Russell 1000, SPDR S&P 500 and Ishares USD Green Bond ETFs showed signs of anti-persistence in returns, indicating a lack of strong relationship or predictability between short-term price movements. The Ishares Global Clean Energy ETF did not reject the random walk hypothesis, suggesting that returns follow a pattern closer to random, where market prices already efficiently reflect all available information. During the conflict, there was a transition in the ETFs' behaviour patterns, as evidenced by the increases in slope values for Ishares MSCI World, Ishares Russell 1000, SPDR S&P 500, Ishares Global Clean Energy and Ishares USD Green Bond. Thus, the possible transition from anti-persistence to long-term memories in ETF returns during the conflict. For portfolio managers, these findings highlight the need to continually adapt investment strategies to manage risks better and take advantage of opportunities in a dynamic and complex investment environment.

Application of Machine Learning Models in Asset Allocation

This study embarks on an exploration of machine learning (ML) models in forecasting the trends of stock indices, with a specific focus on different industries within the Chinese market. Moving beyond the confines of traditional linear regression and standard multi-factor approaches, our research adopts a multi-dimensional analytical framework to decode the complex relationships between various factors and the future returns of industry-specific Exchange-Traded Funds (ETFs) in China. The paper innovatively applies both linear and nonlinear ML models to predict directional shifts in ETF returns, a domain not extensively studied previously. We conduct a thorough comparative analysis of these models, assessing their predictive prowess and dissecting the influence of diverse factors on different industry sectors. This investigation reveals distinct patterns and factor sensitivities unique to each sector, offering new insights into their dynamics. The results are pivotal for asset allocation and investment strategies, as they highlight the nuanced role of ML in financial forecasting. By bridging the gap between traditional financial models and advanced ML techniques, our study presents a novel perspective that enriches the strategic planning in financial markets, especially in the context of the rapidly evolving Chinese economy.

Lever up! An analysis of options trading in leveraged ETFs

AbstractWe examine options trading in leveraged Exchange‐Traded Funds (ETFs) and their impact on the performance of the underlying funds. Using implied volatility innovations in call and put options, we demonstrate that option signals from leveraged ETFs are robust predictors of the underlying ETFs' performance. While both levered and unlevered option signals forecast ETF returns, the levered signal is more pronounced in both magnitude and relevance. This predictivity power primarily stems from inverse leveraged ETFs and during economic downturns. Furthermore, we use the leveraged ETF option signals to develop a trading strategy that produces an average abnormal performance of 1.13% per month.

Retail ETF investing

AbstractUsing marketable order flow data, we analyze key characteristics of aggregate retail exchange‐traded fund (ETF) investing from 2010 to 2021, including allocations, holding period and investment performance. Retail traders allocate 12% more dollar volume to leveraged and inverse ETFs versus nonretail traders. Retail ETF trades distinctly increase with prior ETF returns, in contrast to contrarian stock trading. Estimated ETF holding periods are longer for retail investors versus nonretail. Finally, retail and nonretail ETF trades perform similarly over hypothetical holding periods up to one quarter. Overall, we provide policy‐relevant insights into retail investing behaviours, which have been the subject of recent concern.

EXAMINATION OF THE IMPACTS OF CRYPTOCURRENCY UNCERTAINTY ON EXCHANGE-TRADED FUNDS

In this paper, we use quantile-based analysis and nonlinear conditional autoregressive value-at-risk (CAViaR) model to investigate how cryptocurrency price and policy uncertainty indexes affect exchange traded funds (ETFs) investing in equity securities of companies operating in various disruptive innovation sectors and related to digital assets and blockchain technology. Results of the quantile-on-quantile (QQ) estimates and quantile regression reveal that cryptocurrency price and policy uncertainty generally have a negative impact on the globally traded ETFs suggesting that the changes in cryptocurrency uncertainty lead to destabilizing effects in the ETF market. Moreover, we find that the impact of cryptocurrency uncertainty is asymmetric depending on each quantile of ETF returns even though each ETF is influenced in the same way by each uncertainty index, indicating that cryptocurrency price uncertainty and cryptocurrency policy uncertainty have common effects on the ETF market. A linear quantile causality analysis indicated that there is evidence of a significant causality running from the considered uncertainty indexes to the ETF returns, revealing that cryptocurrency price and policy uncertainty have a significant information value in explaining expected ETF returns. On the other hand, we find that linear quantile models outperform nonlinear CAViaR models in quantiles that are only moderate.

Exchange-traded Funds in India Amid COVID-19 Crisis: An Empirical Analysis of the Performance

Since their inception in 2002, exchange-traded funds (ETFs) in India have made significant progress. The lacklustre record of active investing and the dual benefits of low-cost diversification have elevated the asset class’s credibility. During the COVID-19 crisis, equities ETFs have demonstrated greater resilience than other investment options. Our examination of 35 ETFs that mirror broader indices indicates the performance of Indian ETFs amid the economic shock caused by COVID-19. A comprehensive analysis of the daily NAV, trading price, and benchmark price data from 2015 to the present, post-identification of the pandemic period, indicates three significant inferences: First, tracking errors persist, with significant increases in all three tracking measures throughout the COVID-19 period. Second, during the COVID-19 period, the existence of pricing inefficiencies transitions from discount to premium, with an average daily value of 57.485. Nonetheless, the rate of short-term disequilibrium corrections accelerates dramatically during the COVID-19 phase. Thirdly, it took only eight days during COVID-19 to erase 50% of the disequilibrium between NAV and market price, compared to 170 days before COVID-19. During the COVID-19 era, we also notice a rise in the co-integration between ETF returns and benchmark performance and an increase in the adjustment rate from 18% to 61% per unit of time. The research findings provide portfolio managers and regulators with a comprehensive view of the performance of ETFs before and after an economic shock and offer educational benefits to a growing investor base.

High Frequency Return and Risk Patterns in U.S. Sector ETFs during COVID-19

This study investigates intraday patterns in the eleven sectors of the United States (U.S.). Key contributions are (i) risk and return patterns at specific trading periods on the New York Stock Exchange (NYSE), (ii) whether a specific day return model can predict the next 15-minute positive return, and (iii) the impact of the first vaccination rollout in the U.S. on intraday Exchange-Traded-Funds (ETF) returns. Time-dependent regressions capture risk and return relationships, decision trees in machine learning compare return models, and impulse responses capture the effect of the 2019 coronavirus (COVID-19) vaccine rollout in U.S. 15-minute Standard & Poor’s Depository Receipts (SPDR) Select Sector ETF data is used over 12th March 2020-23rd February 2021. Findings support sector ETF returns fluctuate the most in the first and last 15 minutes. Average returns in the first 15 minutes are the highest, converging to near zero as the trading session continues. Overnight returns contribute the most to volatility. U-shaped patterns into both return and risk exist, especially on Mondays. Mondays and Fridays have the most significant positive returns 15 minutes after the open. Prediction scores using an all-return model were superior to any specific day return model. The first vaccination rollout has a positive effect only in energy, technology, and financial sector ETFs, however with a short-lasting effect on ETFs returns.

Testing for the Long Memory and Multiple Structural Breaks in Consumer ETFs

Abstract This research examines the consumer exchange-traded funds (ETFs) in several industries based on long memory and multiple structural breaks. The autoregressive fractionally integrated moving average (ARFIMA) model indicates that consumer ETF returns in the media, consumer service, food and beverage, and consumer goods industries can be accurately predicted. The autoregressive fractionally integrated moving average and fractionally integrated generalized autoregressive conditional heteroskedasticity (ARFIMA-FIGARCH) model reveals that only the gaming and consumer goods industries have a long memory in volatility. This study establishes that through the iterated cumulative sum square test, multiple structural breaks exist in consumer ETF industries. Results prove that the consumer goods industry has a long memory and multiple structural breaks. Finally, the structural breaks in consumer ETFs have strong asymmetrical effects, indicating that all of the consumer ETF industries are generally unstable. Keywords: The Long Memory, Multiple Structural Breaks, Consumer ETFs, Iterated Cumulative Sums Squares Test.

Are the flows of exchange‐traded funds informative?

AbstractThis paper provides novel evidence of information asymmetry in exchange‐traded fund (ETF) markets. By decomposing daily ETF flows, we find that the unexpected flow component, orthogonal to the components driven by market making and arbitraging, wields substantial power in predicting next day's ETF returns. Informed traders are able to exploit their information advantage to realize an annualized open‐to‐close return of 19.16% or close‐to‐close return of 22.42%. The informativeness of the unexpected ETF component is further confirmed by its strong power of predicting next day's macroeconomic and ETF‐related news, while the market‐making‐ and arbitraging‐driven components are not closely related to forthcoming news.

Price discovery in emerging market ETFs

ABSTRACT This study investigates the price discovery role of exchange-traded funds (ETFs) by examining the predictive relation between the returns of emerging market ETFs traded in the US and the returns to the aggregate equity indices that they track. In a sample that covers 18 countries, we find that ETF returns can predict one-day-ahead returns of their underlying indices. This relation is robust after controlling for the non-synchronicity between markets, serial correlation in index returns, and various determinants of aggregate returns. Moreover, the predictive relation is more pronounced during periods of higher volatility and evidence for bidirectional spillover effects is weak. We also find that an out-of-sample rolling window strategy outperforms investing in the market index several-fold in the majority of the markets, especially in the high-volatility subsample.

Intraday arbitrage between ETFs and their underlying portfolios

Prior research suggests that nonfundamental exchange-traded fund (ETF) price shocks are transmitted to their portfolios through an arbitrage mechanism. We test this proposition by examining minute-by-minute returns and order imbalances but find little evidence that ETF trading impacts underlying returns. Specifically, panel vector autoregression shows that ETF returns do not lead portfolio prices. Instead, arbitrage opportunities arise from order imbalances and price movements in the underlying securities and are subsequently eliminated by ETF quote adjustments, rather than arbitrage trading. We extend our analysis to a daily frequency but still find little relation between ETF trading and constituent security prices.

High Frequency Return and Risk Patterns in U.S. Sector ETFs during COVID-19

This study investigates the presence of intraday patterns in the eleven sectors of the United States (U.S.) economy. Key contributions are in terms of assessing (i) risk and return patterns at specific time periods of the trading session on the New York Stock Exchange (NYSE), (ii) whether a specific day return model can predict the next 15-minute positive return, and (iii) the impact of the first vaccination rollout in the U.S. on intraday Exchange-Traded-Funds (ETF) returns. We use time dependent regressions to capture risk and returns relationships, decision trees in machine learning to compare return models, and impulse responses to capture the effect of the 2019 novel coronavirus (COVID-19) vaccine rollout in the U.S. 15-minute frequency Standard & Poor’s Depository Receipts (SPDR) Select Sector ETF data is used, from 12th March 2020 till 23rd February 2021. Findings suggest that sector ETF returns in both the first and last 15 minutes of the core session fluctuate more than for the rest of the day. Average returns in the first 15 minutes are the highest, converging to near zero as we progress through the day. Overnight returns are the most significant in contributing towards the volatility for any trading session. U-shaped patterns into both return and risk are observed, with Mondays exhibiting a more pronounced U-shaped pattern. Mondays and Fridays have the most significant positive returns 15 minutes after the open. Prediction scores using an all-return model were superior to any specific day return model. The first vaccination rollout has a positive effect only in energy, technology, and financial sector ETFs, however with a short-lasting effect on the ETFs return.

Regime Switching in High-Tech ETFs: Idiosyncratic Volatility and Return

The volatility of asset returns can be classified into market and firm-specific volatility, otherwise known as idiosyncratic volatility. Idiosyncratic volatility is increasing over time with some literature attributing this to the IT revolution. An understanding of the relationship between idiosyncratic risk and return is indeed relevant for idiosyncratic risk pricing and asset allocation, in a context of emerging technologies. The case of high-tech exchange traded funds (ETFs) is especially interesting, since ETFs introduce new noise to the market due to arbitrage activities and high frequency trading. This article examines the relevance of idiosyncratic risk in explaining the return of nine high-tech ETFs. The Markov regime-switching (MRS) methodology for heteroscedastic regimes has been applied. We found that high-tech ETF returns are negatively related to idiosyncratic risk during the high volatility regime and positively related to idiosyncratic risk during the low volatility regime. These results suggest that idiosyncratic volatility matters in high-tech ETF pricing, and that the effects are driven by volatility regimes, leading to changes across them.

Investor attention, ETF returns, and country-specific factors

Using the data of 47 single-country exchange-traded funds (ETFs) traded in the U.S. from 36 countries during 2004–2017, this research examines the impact of investor attention proxied by Google Search Volume Index and home country-specific factors on different quantile of their returns. Evidence first shows that compared with U.S. investor attention, home country investor attention largely correlates with low to medium ETF returns, supporting the attention-induced price pressure hypothesis. Second, home country-specific factors significantly affect ETF returns at different conditional quantiles. We also find that investors prefer investing in a country with strong similarity to that of the U.S., supporting the cross-country information asymmetry hypothesis. Third, an intervening effect of home country-specific factors exists on the relationship between U.S. investor attention and ETF returns. These findings should help government authorities find appropriate strategies to attract foreign investment and upgrade the value of their capital market as well as provide a reference on efficiency for equity investors.

ETFs, Illiquid Assets, and Fire Sales

We document several novel facts about exchange-traded funds (ETFs) holding corporate bonds. First, the portfolio of bonds that are exchanged for new or existing ETF shares (called creation or redemption baskets) often represents a small fraction of ETF holdings – a fact that we call “fractional baskets.” Second, creation and redemption baskets exhibit high turnover. Third, creation (redemption) baskets tend to have longer (shorter) durations and smaller (larger) bid-ask spreads relative to holdings. Lastly, ETFs with fractional baskets exhibit persistent premiums and discounts, which is related to the slow adjustment of NAV returns to ETF returns. We develop a simple model to show that an ETF’s authorized participants (APs) can act as a buffer between the ETF market and the underlying illiquid assets, and help mitigate fire sales. Our findings suggest that ETFs may be more effective in managing illiquid assets than mutual funds.

Risk Mitigation and Return Resilience for High Yield Bond ETFs with ESG Components

The objective of this study is to investigate the effects of ESG components on financial products by using high yield bond exchange traded funds (ETFs). For the objective, we propose a new price correlation and volatility model between an existing financial market such as conventional high yield bond ETFs and a new financial market such as ESG high yield bond ETFs, in which the former affects the latter. Empirical studies show that ESG factors have hedging effects to the downside risk of conventional high yield bond ETF prices in the Covid-19 shocks of March, 2020 from the positive impacts of conventional high yield bond ETF prices on the correlations with ESG ones by using ESG high yield corporate bond ETFs of the Nuveen and the iShares compared with a conventional high yield bond ETF of the SPDR Bloomberg Barclays. Additionally, high yield bond ETFs that take ESG into account are observed to reduce risk compared to a conventional high yield bond ETF. Results also show that ESG high yield bond ETF returns are higher than conventional high yield bond ETF returns in the Covid-19 shocks. The implication from these results is that the value of ESG investing lies in the mitigation of risks and the resilience of expected returns.

Price Discovery in Emerging Market ETFs

Over the last two decades, exchange traded funds (ETFs) have become a preferred investment vehicle to make directional market bets due to their low costs and high liquidity. Moreover, due to the arbitrage activities of authorized participants, prices of ETFs do not deviate materially from the prices of their underlying securities. As a result, ETFs may play a price discovery role such that systematic information is first reflected in these instruments and then transmitted to their underlying securities. We test this hypothesis by investigating the predictive relation between the returns of emerging market ETFs traded in the US and the returns to the emerging market equity indices that they track. In a sample that covers 18 countries, we find that ETF returns can predict one-day-ahead returns of their underlying indices. This relation is robust after controlling for the non-synchronicity between markets, serial correlation in index returns, and various determinants of aggregate returns. Moreover, the predictive relation is stronger during periods of higher volatility. We also find that an out-of-sample rolling window strategy outperforms investing in the market index several-fold in the majority of the markets.

The effects of investor attention and policy uncertainties on cross-border country exchange-traded fund returns

Using a new panel quantile regression approach, this research explores how investor attention and policy uncertainties affect conditional return distributions for 126 cross-border U.S. trading single-country exchange traded funds (ETFs). We find that economic, fiscal, and monetary policy uncertainties negatively impact most of the ETF return quantiles, implying that a non-linear relationship exists, but trade policy uncertainty has asymmetric impacts on current returns that are salient positive in the following period. Therefore, considering four policy uncertainties allows us to completely analyze a single-country ETF market, thus providing valuable insights for ETF portfolio investors, hedge trading, and policymakers.

Unskilled Fund Managers: Replicating Active Fund Performance With Few ETFs

This paper uses Exchange Traded Funds (ETFs) instead of risk factors as benchmarks to examine active mutual fund performance distribution. While transaction costs are included in the ETF returns, that is not true regarding risk factors, making it more challenging to characterize extraordinary performances via alphas. Assessments are based on the estimation of the skilled funds proportion defined by Barras et al. (2010). After evaluating several ETF combinations, we conclude that sets of 3 to 5 ETFs replicate most levels of active fund performance. Finally, we propose specific ETF selection algorithms, whereby we estimate that 95% of active management funds fail to generate value for their investors. Alphas calculated with ETFs are higher than those using risk factors, and the difference is similar to the transaction costs required for investing in risk factor portfolios (Frazzini et al. (2012)).