Next Article in Journal
Economic Policy Uncertainty and Foreign Direct Investment: A Bilateral Perspective on Push and Consistency Effects
Previous Article in Journal
Shadow Economy and the Ecological Footprint Nexus: The Implication of Foreign Direct Investment in ASEAN Countries
Previous Article in Special Issue
Empirical Re-Investigation into the Export-Led Growth Hypothesis (ELGH): Evidence from EAC and SADC Economies
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Are Macroeconomic Variables a Determinant of ETF Flow in South Africa Under Different Economic Conditions?

1
School of Economic Science, North-West University, Gauteng 1174, South Africa
2
Trade Research Entity, North-West University, Gauteng 1174, South Africa
3
College of Business, Al Ain University, Al Ain 64141, United Arab Emirates
*
Author to whom correspondence should be addressed.
Economies 2025, 13(9), 260; https://doi.org/10.3390/economies13090260 (registering DOI)
Submission received: 12 August 2025 / Revised: 28 August 2025 / Accepted: 1 September 2025 / Published: 6 September 2025
(This article belongs to the Special Issue Dynamic Macroeconomics: Methods, Models and Analysis)

Abstract

The objective of this study is to examine the effect of macroeconomic variables on exchange-traded funds (ETFs) returns under different market conditions. The growing prominence of ETFs in emerging markets has over the years drawn much relevance in the academic front for the ability to track the performance of prominent indices, which enhances return perspective. Despite this, ETF returns are influenced by many factors that dampen expected returns; these include macroeconomic variables and changing market conditions. To this extent, monthly data from November 2010 to December 2023 were used in the estimation of the Markov regime-switching model. The findings demonstrate that ETF returns are affected both positively and negatively by macroeconomic factors like inflation, money supply, interest rates, gross domestic product (GDP), and real effective exchange rate. More specifically, the effect tends to vary with market conditions such as bull and bear regimes. This implies there exists adaptive behavior among the ETF market in South Africa, suggesting there are periods of efficiencies and inefficiencies. The findings pose important implications to investors, portfolio managers, and policy makers, all of which is discussed herein.

1. Introduction

The objective of this study is to examine the effect of macroeconomic variables on ETF returns under different market conditions. The study explores the relationship between macroeconomic variables and exchange-traded funds flows, particularly in the context of emerging markets like South Africa. As investors become increasingly interested in ETFs as a flexible investment vehicle, understanding how macroeconomic indicators such as money supply, gross domestic produce (GDP), inflation rates, and interest rates affect these flows under varying economic conditions becomes essential. This inquiry not only sheds light on investor behavior and market dynamics but also provides valuable insights for policymakers and financial analysts aiming to navigate the complexities of the South African economy. By examining the interplay between macroeconomic factors and ETF flows, this study aims to uncover patterns that can inform investment strategies and enhance the stability of the financial market in response to economic shifts. An ETF is a collective investment instrument that consists of a selection of assets designed to mirror the performance of a certain benchmark index. ETFs are essential investment instruments due to their minimal transaction expenses, substantial liquidity, enhanced transparency, and diversification advantages (Wu et al., 2021). Consequently, ETFs represent a cost-effective and readily accessible evolution of mutual funds. The primary distinction between an ETF and a mutual fund is that mutual funds are transacted solely at the conclusion of the trading day at their closing net asset value (NAV), while ETFs are traded continuously during the day at prices dictated by the market (Elton et al., 2019). In recent years, South African ETFs have become increasingly popular as financial instruments. Interestingly, at the end of 2008, the South African ETF market’s combined market capitalization was roughly ZAR 43,699.07. By the end of 2018, however, the market had expanded to a total market capitalization of over ZAR 889, 347.78 (Brown et al., 2021). As a result, during the course of ten years, the capitalization of the South African ETF market more than doubled.
ETFs are a type of pooled investment vehicle because, like index funds, they combine securities from different benchmarks to try and duplicate the performance of a particular benchmark or stock market index (Charteris, 2013). Since ETFs seek to mimic the performance of a certain benchmark, they ought to provide returns that are identical to those of their underlying index.
The total market capitalization of the entire South African ETF industry increased from over ZAR 8.91 billion at the end of 2023 to ZAR 9.62 billion at the end of the first quarter of 2024, a rise of approximately 8%, revealing its current capacity for growth and more so its positive impact on the Johannesburg stock market (Shenjere et al., 2025). Since ETFs aim to replicate the performance of a certain benchmark, they ought to produce returns that are identical to those of their underlying index. Several studies posit that ETF performance is highly predictable due to the low level of tracking error (Gastineau, 2004; Agapova, 2011; Rompotis, 2011). In addition, growth in ETFs in emerging markets has been revealed to amplify their global financial cycle (Converse et al., 2023), giving rise to the need to investigate whether certain economics factors could affect EFT flow in an emerging economy like South Africa. However, it has not been empirically established if ETF flow is influenced by macroeconomic factors in different market conditions in South Africa. The motivation for investigating the possible macroeconomic factors that drive ETF flow under the changing economic conditions in South Africa stem from the increased growth of ETF markets in recent years in addition to the fact that the South African economy has experienced both bullish and bearish periods in recent years.
Recent studies by Ben-David et al. (2014) have examined how ETFs have revolutionized contemporary investment techniques, highlighting how they have made diversification more accessible, while Clifford et al. (2014), in the aim of examining the factors that influence ETF flow, showed that high price/NAV ratios, narrow spreads, and high volume all cause ETF flows to rise, concluding that superior market timing is also essential for ETF flows. Fulkerson et al. (2014, 2017) showed that in predicting the bond-ETF returns, for liquidity provision, premiums and discounts relative to NAV are important elements to consider. However, a much earlier study, namely Kalaycıoğlu (2004), could only establish the correlation and causal relationship between ETF flows and return, thereby only revealing the nature of their relationship. However, in view of these studies, no empirical emphasis has been directed towards revealing the possible macroeconomic drivers of ETFs owing to the fact that macroeconomic policy changes have huge influence on the stock market. Secondly, as an emerging market, the Johannesburg Stock Exchange (JSE) in recent times has grown enough to attract various investments funds, including ETFs (JSE, 2022), hence the great need for this study. The Sygnia Itrix Top 40, Satrix SWIX, SWIZ 40, and the Invest Top 40 are employed as the ETF flow variables in this study, while the effects of macroeconomic variables on ETF return under a changing market conditions (bull and bear) regime is examine using the Markov regime-switching model on the ETF selected variables.
The Markov-regime switching model allows for the parameter to switch in the economic state of the specific country under examination. State-dependent variables are analyzed under the two regimes to follow a first-order Markow process, as shown by the transition probability matrix. These variables include changes in the South African money supply rate (M2), inflation rate (CPI), short- and long-term interest rates (ST-INT and LT-INT), real effective exchange rate (REER) and GDP. Therefore, the aim of this study is to examine the possible drivers of EFT flow in South Africa, as one of the emerging economies in the Sub-Saharan Africa region, under different market conditions. To achieve the aim of this study, monthly return data of the four selected ETF samples were examined against six macroeconomic variables in a regime-switching Markov model, which allows for the macroeconomic variables to switch with the economy’s state. This study differs from other studies on ETFs and macroeconomic variables in its methodological design; for instance, Kittur et al. (2025) investigated the relationship between ETFs and macroeconomic factors using Granger causality and multiple regression, hence revealing the correlation and the corresponding impacts of these the ETFs with macroeconomic factors. Furthermore, this study differs from Nadler and Schmidt (2015), who employed a univariate GARCH model to determine the response and flow of ETFs to macroeconomics announcements; hence, this study significantly differs from other studies in its unique revelation of the different macroeconomic drivers of ETF flow under changing economic conditions since the SA economy has experienced both bullish and bearish economic conditions over the last couple of years. The result of the study convincingly shows that during the bull regime, all the macroeconomic variables—CPI, M2, LT-INT, ST-INT, REER, and GDP—were significantly responsible for flow of the ETF-40 into the South African economy; however, M2, LT-INT, and REER showed a negative impact. Similarly, during the bear condition, all macroeconomic variables were responsible for the flow of STXSWX, with M2 and GDP having positive influence.
Hence, this study contributes to the existing literature by expanding the understanding of ETF dynamics. Thus, it fills the gap in the literature that often prioritizes developed markets, by providing insights into the unique factors that may drive investor behavior in South Africa’s economic landscape. Second, it contributes to the analysis of economic conditions. By examining the impact of different economic conditions, such as periods of economic growth, recession, or stability, on ETF flow, the study enriches the literature on economic cycles and investment patterns. It offers a nuanced perspective on how varying macroeconomic environments can alter investor sentiment and decision-making processes, thereby influencing the attractiveness of ETFs as investment vehicles. Thirdly, this study contributes to policy implications for financial markets. By identifying key macroeconomic determinants of ETF flows, the research provides guidance on how economic policies can be tailored to stabilize or stimulate ETF investments during fluctuating economic conditions. This contribution is crucial for fostering a resilient financial market and promoting sustainable economic growth in South Africa. This study has important implications for stock market policymakers and other stakeholders in the investment markets, such as investors, investment management companies, fund managers, and regulators.
The rest of this paper is structured as follows. Section 2 presents the literature review. Section 3 discusses the study design; Section 4 presents and discusses the results. Section 5 concludes the study.

2. Literature Review

2.1. Theoretical Review

The Adaptive Market Hypothesis (AMH) reframes market behavior as an evolutionary process in which investor psychology, competition among strategies, and institutional structure interact to generate time-varying degrees of market efficiency (Lo, 2004). Under AMH, markets are not permanently efficient or inefficient; instead, efficiency emerges, fades, and re-emerges as participants learn, innovate, and reallocate capital. This makes AMH a natural theoretical lens for studying whether macroeconomic variables determine ETF flows in South Africa differently across bullish and bearish regimes: the sensitivity of flows to macro shocks is expected to depend on the prevailing ecological state of the market.
AMH identifies three core mechanisms relevant to ETF flows. First, bounded rationality and heuristic learning mean investors adopt simple rules (momentum, yield-chasing, hedging, etc.) that perform variably across states; such heuristics can amplify inflows in bullish phases and sharpen outflows in bearish phases (Lo, 2004). Second, natural selection of strategies implies that profitable strategies attract capital until excess crowding and adaptation reduce their efficacy, so the same macro signal (e.g., inflation surprise) may produce different flow responses over time. Third, institutional evolution changes in ETF availability, trading costs, and investor composition alters transmission channels from macro fundamentals to flows, particularly in an emerging market context like South Africa, where product penetration and investor types have evolved rapidly (SARB, 2020). Empirical AMH work for emerging markets shows global push factors (VIX, US rates) exert larger effects during stress, consistent with AMH’s state-dependent predictions (Arslanalp et al., 2020; Bouri et al., 2021). South Africa-specific studies similarly document asymmetric flow responses to shocks (Shenjere et al., 2025).
Operationalizing AMH for ETF-flow research requires distinguishing periods of efficiency and inefficiency. Under AMH, these are not binary labels but degrees: an “efficient” period is one in which returns (or flows) exhibit low short-horizon predictability and weak exploitable autocorrelations, while an “inefficient” period displays statistically significant predictability, autocorrelation, or persistent pricing/flow anomalies that permit above-average risk-adjusted profits until arbitrage or learning dissipates them (Lo, 2004). Empirically, researchers identify such periods using rolling or time-varying tests and models: rolling variance ratio and serial-correlation tests (Lo & MacKinlay, 1988) to detect short-term predictability; Hurst exponent and runs tests for persistence; Bai–Perron structural-break tests to locate regime shifts (Bai & Perron, 1998); Markov switching models to classify discrete bull/bear states (Hamilton, 1989); and time-varying parameter (TVP) or TVP-VAR frameworks to trace continuous evolution in betas linking macro variables to flows (Koop & Korobilis, 2013; Lubik & Matthes, 2015). Local projections with interaction terms (macro × regime dummy) are a complementary approach to estimate asymmetric impulse responses.
Conclusively, AMH predicts that macroeconomic determinants of South African ETF flows will be regime-dependent and evolving: the same CPI, M2, LT-INT, ST-INT, REER, and GDP shock can elicit distinct flow patterns in bullish versus bearish periods because investor heuristics, liquidity premia, and institutional settings adapt. Testing these predictions requires econometric tools that allow coefficients and predictability to change over time or across discrete regimes, thereby aligning the empirical design with AMH’s evolutionary logic.

2.2. Empirical Review

As ETFs have grown in popularity as an investment instrument worldwide (Aggarwal & Schofield, 2014), the relationship between macroeconomic factors and ETF flows has drawn more attention in the financial literature. It is well recognized that macroeconomic variables such as GDP growth, unemployment, inflation, and interest rates affect market dynamics and investor behavior. Given that economic conditions in developing nations can be unstable and diverge greatly from those in established economies (Arayssi, 2020), it is especially important to comprehend how these factors affect ETF flows. Research in developed markets has established that macroeconomic indicators could serve as determinants of investment flows. For instance, Koepke (2018) revealed that rising interest rates may lead to reduced ETF inflows as investors seek safer assets, while strong economic growth can boost investor confidence and drive capital into equity ETFs. However, there is a scarcity of literature that specifically examines these relationships within the context of South Africa, an economy characterized by its unique challenges and opportunities.
Very little scholarly work focuses solely on the factors that influence ETFs. From the research that exists, the majority concentrates on the ETF market quality for ETFs (Box et al., 2019; Lerner, 2022; Höfler et al., 2023) or the competition between index ETFs and index mutual funds (Narend, 2014; Blitz & Vidojevic, 2019; Lesmeister et al., 2022). Limited empirical research relates the drivers or determinants of fund flows to ETFs, hence the essentiality of this study in the South African context. South Africa’s economic landscape is marked by high levels of inequality, fluctuating commodity prices, and varying levels of political stability (Levy et al., 2021), which can all impact investor sentiment and behavior. As such, the response of ETF flows to macroeconomic variables may differ significantly compared to more stable economies. Previous studies such as Hoang (2018) have highlighted the importance of understanding local economic conditions and investor psychology when analyzing investment flows in emerging markets. Moreover, the role of macroeconomic variables may vary under different economic conditions such as periods of economic expansion versus recession—making it essential to explore the dynamic interplay between these factors. By investigating how macroeconomic variables influence ETF flows in South Africa under varying economic conditions, this study aims at bringing comprehensive understanding of investment behavior in emerging markets, filling a critical gap in the existing literature. As shown by Lee and Swaminathan (2000), historical trading volume can play a significant role in explaining the trading methods that investors use to trade ETF equities. Therefore, an ETF’s historical trading volume ought to be connected to both fund flows and investor trading patterns.
Using a sample of index and financial ETF trading in the United States (U.S.), Chiu et al. (2012) investigated the relationship between funding liquidity and equity liquidity during the subprime crisis era using ETFs. They investigated a very basic regression model that links a funding liquidity variable and traditional drivers to an ETF’s liquidity. They discovered that improving funding liquidity can also boost equity liquidity, albeit the impact on financial ETFs is larger than that on index ETFs. However, their study fails to include the underlying index liquidity, neglecting the redemption process and treating ETFs as if they were mere equities. The enhancement in liquidity shown in Chiu et al. (2012) may actually stem from an increased liquidity of the underlying stock basket rather than only from a direct reduction in the funding constraints encountered by liquidity providers.
The ETFs has increased from tens of billions of dollars in 2000 to almost a trillion dollars in 2010 in the U.S., which has led to massive interest from investors. To determine the factors that influence the movement of ETFs, according to Clifford et al. (2014), ETF flows rise in response to high price/NAV ratios, modest spreads, and high volume. The authors also found minimal indication of better market timing in ETF flows using a dataset of more than 500 ETFs from 2001 to 2010. Moreover, they proved that investors’ naïve extrapolation bias is more likely to be the cause of return chasing in mutual funds and ETFs, which has fueled the expansion of the ETF market.
Since retail investors’ investment decisions should be influenced by industry-wide categorization, as demonstrated by Jame and Tong (2014), ETF activity should also be determined by relative metrics of performance1. Barber et al. (2005) discovered that the Capital Asset Pricing Model (CAPM) based alpha better predicts fund flows than either the Carhart four-factor alpha or the Fama–French three-factor alpha, demonstrating the value of simple benchmarking and the application of beta. According to M. Baker et al. (2011), investors’ irrational preference for high volatility causes low-beta equities to perform better than high-beta companies, suggesting that demand for sector ETFs rises with beta. Additionally, they contended that institutional investors are deterred from taking advantage of this low-risk oddity by benchmarking since doing so would expose them to the danger of losing money relative to their benchmarks when they invest in low-beta stocks. Frazzini and Pedersen (2014) provided an explanation for the low-risk paradox by stating that investors’ margin limitations drive them to bid up high-beta assets, which results in low alphas for high-beta companies2.
H. K. Baker et al. (2015) examined the low-risk anomaly by breaking it down into micro and macro impacts, with a strong focus on sector ETFs. They discovered that a sizable portion of the outperformance of low-beta equities is determined by industry betas. The study revealed that investors predictably enhance their market exposure by altering their exposure to other sectors, as the findings indicated that macro-effects drive the returns of anomaly. Consequently, sector ETFs in particular should see an increase in demand as beta.
To assess whether volatility transfers from ETFs to their primary underlying stocks and, if so, to identify the factors influencing the magnitude of these spillovers, Krause et al. (2014) established this approach by identifying probable drivers of volatility spillover for equities included in ETFs using Kyle’s (1985) linear pricing paradigm. The authors used daily ETF and component stock return and price data from 1 March 2003 to 31 December 2013, representing all of the sectors in the S&P 500 market. The outcome demonstrated the economic impact of volatility spillover from ETFs to their largest component stock. The proportion of each stock held by the funds, variances from NAV, ETF flow of funds, and ETF market capitalization all reflected an increase in liquidity as a result of these spillovers.
With the intention of examining how beta and a broader range of risk factors influence ETF flow and trading of ETF sectors, Peltomäki (2017) used the returns of the ETF sector from 31 May 2006 to 6 March 2012 in a multivariate analysis. The findings show a diminishing (rising) and U-shaped (inverse U-shaped) relationship between beta and ETF trading (ETF flow). While significant ETF-related risk drives ETF trading, great relative performance and investor attention increase ETF flow. The findings imply that unusual investment decisions, such as reallocation decisions to investment funds, are influenced by distinct causes of trading activity. This analysis also demonstrates that beta plays a significant role in understanding investor behavior in ETFs, with different effects seen for ETF flow and trading.
Conclusively, the literature has not been able to extensively establish the importance of macroeconomic variables as key determinants of ETF flows in South Africa, particularly under varying economic conditions. These variables, for example, exchange rates, inflation, interest rates, and economic growth, directly influence investor sentiment, risk appetite, and market performance, shaping the movement of capital into and out of ETFs. Understanding these macroeconomic factors is essential for investors looking to maximize portfolio allocation and for policymakers hoping to improve market stability, given South Africa’s susceptibility to both internal and international economic swings. Thus, it is crucial to analyze how macroeconomic variables and ETF flows interact in order to predict market trends and make wise investment choices.
Moreover, to the knowledge of the authors, no existing study has established the real macroeconomic factors that affect ETF flow, especially in changing economic conditions, hence the need for this study. Considering the fact that real economic activities are impacted by macroeconomic indices (Nakajima, 2023) and considering the fact that the South African economy has experienced bullish and bearish cycles for years (Laubscher, 2020), this study seeks to establish specific drivers of ETF flows in one of the world’s emerging economies. Our study stands out in a unique way by investigating the macroeconomic level factors that could be determinants of ETFs even under changing market conditions. This would go a long way in equipping investors and policymakers with the necessary monetary policies that could be monitored while investing in ETFs.

2.3. Research Hypotheses

2.3.1. Consumer Price Index (CPI/Inflation)

Inflation is a critical macroeconomic factor influencing capital allocation. High inflation erodes purchasing power and creates uncertainty, reducing investor appetite for risky assets such as equities while increasing demand for inflation-hedging instruments (Bekaert & Wang, 2010). In the South African context, elevated CPI has historically driven volatility in financial markets and shaped investor flows (SARB, 2020). Under the AMH, investors adjust differently depending on regimes: in bullish markets, moderate inflation may be tolerated as a signal of growth, whereas in bearish markets, high inflation is likely to intensify ETF outflows due to risk aversion. On this basis the following hypothesis is derived:
H1. 
CPI has a significant and regime-dependent effect on ETF flows in South Africa, with higher inflation reducing equity ETF flows more strongly in bearish markets than in bullish markets.

2.3.2. Broad Money Supply (M2)

M2 reflects liquidity conditions in the economy. Expansions in money supply often stimulate asset prices by improving credit availability and supporting investment demand (Chen et al., 1986). In bullish conditions, higher M2 growth may fuel ETF inflows, as liquidity supports risk taking. In contrast, during bearish conditions, rapid monetary expansion can be perceived as inflationary or destabilizing, thereby discouraging flows (World Bank, 2020). To this extent, we predict the following hypothesis:
H2. 
Growth in M2 has a positive impact on ETF flows in bullish conditions but a weaker or negative effect in bearish conditions.

2.3.3. Long-Term Interest Rates (LT-INT)

Long-term interest rates determine discount rates for equities and reflect expectations of inflation and growth. Rising long-term yields tend to depress equity valuations but can enhance the attractiveness of fixed-income ETFs (Fama & French, 1989). In expansions, higher long-term yields may attract inflows to bond ETFs, while in contractions, they amplify outflows by raising borrowing costs and reducing market confidence. Accordingly, the following hypothesis is provided:
H3. 
Long-term interest rates significantly influence ETF flows, with positive effects on ETF flows in bullish markets but negative effects on ETF flows in bearish markets.

2.3.4. Short-Term Interest Rates (ST-INT)

Short-term interest rates mirror central bank monetary policy. Rate increases can attract foreign capital into bond or money market ETFs, but they simultaneously reduce equity ETF flows by tightening liquidity and raising opportunity costs (Laopodis, 2010). Under bearish conditions, rate hikes are particularly damaging to ETF flows, as they worsen liquidity constraints. Thus, we hypothesize the following effect:
H4. 
Short-term interest rates positively affect fixed-income ETF flows but reduce equity ETF flows, with stronger negative effects in bearish conditions.

2.3.5. Real Effective Exchange Rate (REER)

The REER captures currency competitiveness relative to trading partners. Depreciation can improve export competitiveness, benefiting equity markets, but may also deter investors if perceived as a sign of instability (Badinger & Reuter, 2017). Under AMH, the impact is regime-dependent: in bullish markets, depreciation may attract inflows into exporter-oriented ETFs, while in bearish markets, it triggers outflows as investors retreat from perceived risk. Hence, the following is hypothesized:
H5. 
REER movements have asymmetric effects on ETF flows, with depreciation supporting flows during bullish markets but discouraging flows during bearish markets.

2.3.6. Gross Domestic Product (GDP)

GDP growth is a broad measure of economic strength and investor confidence. Higher growth prospects encourage capital inflows into equity ETFs, while weak growth dampens investor sentiment (Rapach et al., 2013). Under AMH, these effects are stronger in bullish regimes, when optimism dominates, and weaker in bearish regimes, where risk aversion constrains capital inflows. On this basis, the following hypothesis is derived:
H6. 
GDP growth has a positive effect on ETF flows and is stronger in bullish conditions than in bearish conditions.

3. Data and Methodology

3.1. Data Collection and Sampling

This study analyzes monthly data spanning from November 2010 to December 2023. The sample period is determined by data availability, as the Invest Top 40 and Invest SWIX 40 ETFs have been available only since October 2010. The dependent variables include four ETF indices: Invest Top 40, Invest SWIX 40, Satrix SWIX Top 40, and Sygnia Itrix Top 40—following the selection in Kunjal et al. (2021). Moodley et al. (2022) suggests that macroeconomic variables—such as inflation, money supply, short- and long-term interest rates, GDP, and the real effective exchange rate—affect stock market returns either positively or negatively. Islamiah (2025) further argued that central bank interventions in monetary policy, impact stock prices by altering investor purchasing power. These changes influence the demand and supply of ETFs, which in turn affect their pricing and return potential (Hopman, 2007). Accordingly, this study includes the mentioned macroeconomic variables and transforms them into growth rates. The ETF data were obtained from Bloomberg terminal, whereas the macroeconomic variables were obtained from the South African Reserve Bank (SARB). It is important to note that GDP was converted from quarterly figures to monthly figures using the quadratic average interpolation method in EViews 14, as advocated by Dlamini (2017). A summary of all variables used is presented in Table 1.

3.2. Model Specification

To assess how macroeconomic factors influence JSE ETF returns amid shifting market conditions, this study considers a regime-switching model. More specifically, the Markov regime-switching model of conditional mean with constant transition probabilities is utilized, as it’s the preferred model among academics when examining market conditions in South Africa (see Moodley et al., 2024; Mofokeng & Moodley, 2025; Nkomo & Moodley, 2025). Furthermore, the model presents an added advantage such that it allows for exogenous changes over specific regimes, whereas other nonlinear models only consider endogenous switching. Accordingly, the model specification is
Ι t = μ c t + α 0 i c t Δ C P I + α 1 i c t Δ M 2 + α 2 i c t Δ S T I N T + α 3 i c t Δ L T I N T +   α 4 i c t Δ G D P                                           + α 5 i c t Δ R E E R + ε c t ,
The variable Ι t represents the returns on JSE ETFs. The intercept μ c t  reflects the regime-specific mean. The state variable ct takes on two values: 1 and 2, representing bull and bear market regimes, respectively. Each macroeconomic variable in the model has coefficients that depend on the prevailing regime. These variables include the following:
  • ΔCPI: inflation rate growth rate;
  • ΔM2: money supply growth rate;
  • ΔST_INT: short-term interest growth rate;
  • ΔLT_INT: long-term interest growth rate;
  • ΔGDP: gross domestic product growth rate;
  • ΔREER: real effective exchange growth rate.
Each regime is assumed to follow a first-order Markov process, characterized by a transition probability matrix. In this framework, the likelihood of transitioning to a particular regime depends solely on the most recent state rather than the full history of past states. This dependency structure is represented as follows:
P r o b C t 1 = i = P r o b i j t
In this context, i j represents the probability of transitioning from regime i at time t − 1 to regime j at time t. These transition probabilities are assumed to remain constant over time such that Prob(t) = Probij.

4. Empirical Results

4.1. Descriptive Statistics

Table 2 presents the descriptive statistics, variance inflation factor (VIF) test results, and the outcomes of unit root and stationarity tests. Panel A reveals that the ETFT40 ETF exhibits the highest average return, followed by STX40, STXSWX, and ETFSWX. In terms of maximum returns, ETFSWX records the highest, while STX40 shows the lowest. When examining minimum returns, STXSWX has the lowest value, followed by ETFSWX, STX40, and ETFT40. Although ETFSWX achieves the highest maximum return, it also displays the highest standard deviation among the ETFs, indicating greater volatility. This higher standard deviation reflects a wider spread between its minimum and maximum returns, suggesting more frequent and significant fluctuations over the sample period. Skewness analysis shows that the returns of ETFSWX and STXSWX are negatively skewed, meaning the mean is located to the left of the median. Conversely, ETFT40 and STX40 returns are positively skewed, indicating that the mean lies to the right of the median. These observations are further supported by the kurtosis values, all of which exceed 3, implying that the return distributions are leptokurtic and deviate from the normal bell-shaped curve. This non-normality is confirmed by the Jarque–Bera test results.
Table 2 also reports the descriptive statistics for the associated macroeconomic variables. As shown in Panel A, the money supply exhibits the highest average growth rate, while the real effective exchange rate shows the lowest. In contrast, GDP records the highest maximum growth rate, whereas the money supply has the lowest maximum value. These results suggest that although the money supply grows consistently on average, it does not experience the most substantial growth spikes within the sample period. Interestingly, the short-term interest rate shows the highest maximum growth value but also has the largest standard deviation among the macroeconomic variables. This indicates that short-term interest rates are the most volatile or risky, as evidenced by the wide gap between their maximum and minimum values. Furthermore, all macroeconomic variables—except the short-term interest rate—exhibit positive skewness, implying that their distributions have longer right tails and that their means are to the right of the medians. These findings indicate that the distributions of the macroeconomic variable growth rates deviate from normality. This is confirmed by the Jarque–Bera test and further supported by the kurtosis values, all of which exceed 3—except in the case of the real effective exchange rate—indicating leptokurtic distributions.
Panel B reports the results of the VIF test. The VIF values for all macroeconomic variables are close to 1, indicating a low level of multicollinearity among the explanatory variables. This suggests that their inclusion in the model is appropriate and does not pose multicollinearity concerns, thereby strengthening the robustness of the model specification. Panel C presents the results of the augmented Dickey–Fuller (ADF) unit root test. The ADF test statistics are more negative than the corresponding critical values at the 5% significance level, leading to the rejection of the null hypothesis of a unit root. This implies that both the ETF returns and the macroeconomic variables are stationary. These findings are further supported by the Kwiatkowski–Phillips–Schmidt–Shin (KPSS) test, which fails to reject the null hypothesis of stationarity, providing additional confirmation. Moreover, the ADF break-point test—accounting for potential structural breaks in the data—yields consistent results. The ADF statistics remain more negative than the 5% critical value, reinforcing the conclusion that the time series are stationary, even in the presence of structural changes.

4.2. Unconditional Correlation Results

Table 3 presents the results of the unconditional correlation analysis between ETF returns and macroeconomic variables. The findings reveal that the growth rate of inflation has a significantly negative impact on the returns of ETFSWX, STX40, and STXSWX. Similarly, the growth rate of money supply negatively influences the returns of ETFT40, STX40, and STXSWX. Both the short-term interest rate growth and GDP growth exhibit a significant negative relationship with all ETF returns, with the exception of STX40. Additionally, the long-term interest rate growth shows a significant negative effect across all ETF returns. In contrast, the growth rate of the real effective exchange rate negatively affects only the returns of ETFSWX and STX40. These results suggest that fluctuations in macroeconomic indicators generally have a negative impact on ETF performance, indicating that increases in these macroeconomic variables are associated with declines in ETF returns. The findings also imply a linear relationship between macroeconomic variables and ETF returns. However, the analysis does not account for possible nonlinear dynamics. To address this limitation and capture regime-dependent behavior, the study proceeds with the estimation of a Markov regime-switching model.

4.3. Empirical Model Results

4.3.1. Markov Regime-Switching Results

The Markov regime-switching results is presented in Table 4 below. In Panel A, the average returns for all ETF returns are significant and positive in a bull regime. However, the ETFSWX has the highest average return, followed by ETF40, STXSWX, and STX40 returns. The findings suggest that on average, when the South African market is in a bullish state, the EFT returns will increase, permitting favorable investment gains. These findings are expected, as in a bull market conditions, returns increase over time and would be positive, as found in this study (Davies, 2013). The associated volatility in regime 2 is higher than the associated volatility in regime 1. Thus, regime 1 is the bullish market condition, whereas regime 2 is the bearish market conditions. More specifically, the volatility of all ETFs is negative and significant in a bear regime. However, STXSWX has the highest volatility, followed by STX40, ETFSWX, and ETFT40. Despite ETFSWX attaining the highest average return, the volatility is in the lower range as compared to other ETF returns. This implies that the investment gains are attainable with a limited amount of risk, and such findings are favorable for the risk adverse investor wanting to mitigate their losses and simultaneously acquire enhanced returns.
In Panel B, in the bear market condition, the ETFSWX and ETFT40 average returns are negative and significant, but the STX40 and STXSWX average returns are positive and significant. The former suggests that the returns are decreasing in a bear market condition, whereas the latter suggests that the returns are increasing in a bear market condition. These findings of alternative effects do not come as a shock because AMH advocates for such findings. Moreover, ETF indices track the performance of larger indices, and such indices tend to perform much better during market uncertainty. Similarly, in a bear regime, the volatility associated with all ETF returns is negative and significant and much higher than the associated volatility in bull market conditions. This implies that the bear market conditions are more volatile, and such is expected due to the characteristics associated with bear markets.
Under bull market conditions, the inflation growth rate has a significantly positive impact on the returns of ETFSWX and ETFT40 but a significantly negative effect on STX40 and STXSWX. The money supply growth rate negatively influences ETFT40 returns, while the short-term interest rate growth positively affects the returns of ETFT40 and STXSWX. Both the long-term interest rate growth and the real effective exchange rate growth exert a negative effect on the returns of all ETFs. Conversely, GDP growth has a significantly positive impact on the returns of ETFT40 and STXSWX.
In bear market conditions, inflation growth remains positively associated with ETFT40 returns but has a significantly negative effect on STXSWX returns. The money supply growth rate shows a significantly positive relationship with all ETF returns except ETFT40. Short-term interest rate growth positively affects STX40 returns but negatively impacts STXSWX. Long-term interest rate growth continues to negatively influence all ETF returns, with the exception of ETFSWX. GDP growth positively affects STX40 and STXSWX returns, while it negatively impacts ETFSWX. Similarly, real effective exchange growth has a negative effect on the returns of STX40 and STXSWX
Panel C presents the associated diagnostic tests to further validate the robustness of the model. The results show that the DW-STAT statistic equals 2 for all the regressed models, indicating that there is no autocorrelation in the residuals. Furthermore, the LM-Stat is insignificant, as evidenced by the p-values. Therefore, the study fails to reject the null hypothesis of no autocorrelation in the model residuals. These results suggest that the model is free from robustness issues and is suitable for the analysis.

4.3.2. Transition Probabilities and Expected Duration Results

Table 5 presents the transition probabilities and expected durations associated with each ETF return. In Panel A, it is evident that the transition probabilities for the returns of ETFSWX (0.950437), STX40 (0.933164), and STXSWX (0.810897) in a bull market regime are higher than those in a bear regime, where the probabilities are 0.766983, 0.156015, and 9.30 × 10−9, respectively. These findings suggest that the returns of ETFSWX, STX40, and STXSWX tend to remain in a bull regime for a longer duration than in a bear regime. This is further supported by the expected duration results, where in the bull regime, ETFSWX (20.17642 months), STX40 (14.96205 months), and STXSWX (5.28812 months) all show longer durations compared to their respective durations in the bear regime, which are 4.291526 months, 1.184855 months, and 1 month, respectively. Conversely, the transition probability for ETFT40 returns in a bull regime (0.136046) is lower than in the bear regime (0.857261), indicating that ETFT40 returns stay longer in a bear regime than in a bull regime. This is also confirmed by the expected duration, which is higher in the bear regime (7.005816 months) compared to the bull regime (1.157469 months). Overall, these findings suggest that bull market conditions tend to dominate the ETF returns. This result aligns with the study of Kunjal (2022), which also found bull market conditions to be more persistent in ETF returns.

4.3.3. Smooth Regime Probabilities Results

Figure 1 below presents the smooth transition probabilities associated with the ETF returns. If one takes note of the trajectory of the ETF returns, it is evident that the returns enter and exit in the bull and bear market conditions at different periods in the sample. However, the ETFSWX and STX40 are less persistent in the switching among regimes, as we can see that the ETFSWX returns stay for prolonged periods in a bull regime, whereas STX40 returns stay for prolonged periods in a bear regime. These findings are supported by the transition probabilities, as it was found that the bull and bear market condition was dominant for these ETF returns. Conversely, the returns of the ETFT40 and STXSWX are more persistent such that when the returns enter a bull regime, they do not stay for prolonged periods in the market conditions, as they switch almost instantly into the bear regime. These findings are further supported by the associated transition probabilities. If one takes note of the periods of bearish conditions, it is evident that all ETF returns were bearish during 2019, and these findings coincide with the COVID-19 pandemic, as ETF returns fell and maintained bearish returns during such a period. However, we see that post COVID-19, the ETFSWX returns recovered from the pandemic, as the returns entered a bullish state and remained in a bullish state for the remainder of the sample. However, the remainder of the ETF returns entered and exited the bullish and bearish periods, implying that a full recovery from the pandemic is not evident. Thus, the ETFSWX returns are a more robust investment, as they are able to mitigate financial market uncertainty and generate positive returns.

5. Discussion of Results

It is important to note from the onset of this section that the empirical literature surrounding the current study is limited, with not many academics attempting to examine the desired effect of macroeconomic variables on ETF returns during changing market conditions. To this extent, it is near impossible to critically evaluate the current literature in relation to the findings of this study. Accordingly, the study focuses on analyzing the findings from the perspective of market participants.
It is evident from the findings that macroeconomic variables have an alternating effect on ETF returns under changing market conditions. That is, when the market is in a bullish state, inflation is found to positively influence ETFSWX and ETFT40 index returns but negatively influence STX40 index returns. These findings suggest that when markets are stable, and monetary policy adjustments are conducted, investors should limit their investment strategies to the ETFSWX and ETFT40 index returns, as changes in inflation will enhance returns. On the other hand, investors should limit investing in STX40 index returns during a bull market condition, as it will reduce the returns perspective as inflation targeting is conducted. Therefore, STX40 index returns are more sensitive to changes in inflation as opposed to ETFSWX and ETFT40 index returns. However, in bear market conditions, the STX40 index returns are positively influenced by inflation. This implies when volatile market conditions arise in the South Africa market, investors should incorporate STX40 index returns in their portfolio, as this will yield positive returns despite changes in the inflation rate.
Similarly, in a bull market condition, investors should not consider any ETF index return in their investment strategies or portfolios, as ETF index returns are negatively influenced by money supply. However, in a bear regime, investors must factor in ETFSWX, STX40, and STXSWX index returns as these are resilient to volatile market conditions and changes in money supply. Moreover, when long-term interest rates and real effective exchange rates fluctuate, investors should not consider any of the ETF index returns, as these will negatively influence return perspective irrespective of the market condition. In a bull market condition, investors should only consider ETFSWX, ETFT40, and STXSWX index returns, as changes in short-term interest rate and GDP will enhance returns, thereby ensuring favorable returns for investors. However, in the bear market condition, investors should include STX40 returns in their portfolios or investment strategies, as this will ensure positive returns despite volatile market conditions and monetary policy announcements to GDP and short-term interest rates.
The findings further demonstrate that ETF returns are characterized as bullish over the sample period, suggesting that the returns are increasing for prolonged periods of time. These findings coincide with a study by Kunjal et al. (2021), where the academics also found that bullish market conditions are persistent among ETF indices returns that track the JSE Top 40 index. Accordingly, there exists evidence that the ETF indices returns are resilient to volatile market conditions, suggesting that passive investment strategies should include investments in the indices that track the JSE Top 40 and JSE All-share index. Moreover, the findings demonstrate the resilient nature of South African ETFs in the context of passive investing, as this is a much safer strategy to mitigate the volatile conditions of the South African financial market and ensure prolonged periods of positive returns.
In addition to the above findings, it was found that the South African ETF market expresses adaptive market behavior, as highlighted in the AMH of Lo (2004). According to AMH, for there to be adaptive behavior, the effect of macroeconomic variable on ETF returns must be time-varying and regime-specific. This implies that the effect must alternate with each market condition, which is evident in this study. Thus, there exist periods of alternating efficiency, where the South African ETF market is efficient and inefficient. These findings are further supported by Kunjal et al. (2021), Kunjal (2022) and Shenjere et al. (2025); the academics found adaptive market behavior demonstrated in the South African ETF market. In summary, these findings suggest that investors can earn excess returns in the South African ETF market by using past information, such as prices, to make informed decisions. These findings contradict the notion proposed by the Efficient Market Hypothesis (EMH), in which markets are efficient, and as such, excess returns are not possible.
Collectively, these findings have serious implications for both investors and policy makers. For example, macroeconomic variables influence ETF returns. However, the effect varies with the state of the market (bull or bear market conditions), risk factor (type of macroeconomic variables), and change in risk factor (increase or decrease in the macroeconomic variable). Thus, when investors devise investment objectives, they must consider these changes, as it will contribute to portfolio return enhancement or decline. Moreover, policymakers must devise strategies through policy implications to ensure there is stability within the ETF market. This is because when regulators revise monetary policy, it has adverse effects on the ETF market, which creates instability and causes panic among investors. Therefore, there must be a balance between monetary policy control and market quality.

6. Conclusions

This study aimed to examine the influence of macroeconomic variables on ETF returns under different market regimes. To accomplish this, monthly data from November 2010 to December 2023 were analyzed. The dependent variables consisted of four ETFs tracking the TSE/JSE Top 40 index: Invest Top 40, Invest SWIX 40, Satrix SWIX Top 40, and Sygnia Itrix Top 40. The explanatory variables included six key macroeconomic indicators: inflation growth rate, money supply growth rate, short-term interest rate growth, long-term interest rate growth, GDP growth rate, and real effective exchange rate growth rate.
Findings from the Markov regime-switching model revealed that the effects of macroeconomic variables on ETF returns are regime-dependent, varying significantly between bull and bear markets. For example, inflation growth had a significantly positive effect on the returns of ETFSWX, ETFT40, and STXSWX in bull markets, but its influence turned significantly negative during bear markets. Similarly, money supply growth negatively affected all ETF returns during bull markets yet exhibited a significantly positive effect on STX40 and STXSWX returns in bear markets. The short-term interest rate growth had a significantly positive impact on all ETF returns except STX40 in bull markets, but its effect became negative during bearish phases. Long-term interest rate growth consistently exerted a negative influence across all ETFs in bull markets, with a continued negative effect on ETFT40, STX40, and STXSWX during bear markets. Comparable regime-dependent patterns were also observed for GDP growth and real effective exchange growth rates, highlighting the dynamic relationship between macroeconomic conditions and ETF performance.
Despite the study contributing to the existing literature, this study is not without limitations. For instance, the study isolated the independent variables to six macroeconomic variables. Although this was performed in accordance with the empirical literature, future studies can enhance the range of macroeconomic variables by considering policy uncertainty, different forms of money supply like M1 and M3, unemployment rate, and exchange rates. Furthermore, the study only considered ETFs that track the JSE Top 40 index, but future research can extend the analysis to examine different ETFs that track the various sectors on the JSE, such as the financials, resources, and industrial sectors. Moreover, this study focuses on South Africa as an emerging market; future studies can enhance the analysis across African emerging markets, which will provide a more holistic view and yield added benefits for international and domestic investors.
In summary, the findings of the study provide interesting areas of enhancement of the emerging market literature. The study introduces a new area of research in emerging markets, which is dominated by developed market literature, by looking at how macroeconomic variables influence ETF returns during alternating market conditions. Consequently, new propositions on economic cycles and investment patterns are introduced to enhance the resilience of emerging markets. Moreover, the study factors in the market efficiency of ETF markets in the emerging market setting, which had yet to be done before. Therefore, the study introduces a new concept of adaptive market behavior to the emerging market literature, which before now was non-existent.

Author Contributions

Conceptualization, F.M.; methodology, F.M.; software, F.M.; validation, F.M.; formal analysis, F.M.; investigation, F.M.; resources, F.M.; data curation, F.M.; writing—original draft preparation, F.M. and B.L.; writing—review and editing, F.M., B.L. and M.I.T.; visualization, F.M., B.L. and M.I.T.; supervision, F.M.; project administration, F.M., B.L. and M.I.T. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data are available on request from the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviations

ADFAugmented Dickey–Fuller
AMHAdaptive market hypothesis
CAPMCapital asset pricing model.
CPIConsumer price index
ETFExchange traded funds
ETFSWXInvest SWIX 40 ETF
ETFT40Invest Top 40 ETF
GDPGross domestic product
JSEJohannesburg stock exchange
KPSSKwiatkowski–Phillips–Schmidt–Shin
LT-INTLong-term interest rate
M2Money supply
NAVNet asset value
REERReal effective exchange rate
ST-INTShort-term interest rate
STX40SWIX Top 40 ETF
U.S.United States
VIFVariance inflation factor

Notes

1
It should be noted that relating return chasing behaviour in sector ETFs to only behavioural biases is not unproblematic since return chasing behaviour in investing in passively managed sector ETFs can also be a rational active strategy for investors given the evidence for the profitability of industry momentum strategies (e.g., Moskowitz & Grinblatt, 1999).
2
In a broader overview of the explanations for the low-volatility anomaly, Blitz et al. (2019) relate it to the assumptions of the CAPM from different aspects and consider leverage constraints and short selling restrictions as reasons for the existence of the anomaly.

References

  1. Agapova, A. (2011). Conventional mutual index funds versus exchange-traded funds. Journal of Financial Markets, 14(2), 323–343. [Google Scholar] [CrossRef]
  2. Aggarwal, R., & Schofield, L. (2014). The growth of global ETFs and regulatory challenges. In Advances in financial economics (pp. 77–102). Emerald Group Publishing Limited. [Google Scholar]
  3. Arayssi, D. (2020). The developing countries: Challenges in the global economy. Available online: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3589489 (accessed on 6 March 2025).
  4. Arslanalp, S., Claessens, S., & Puy, D. (2020). Drivers of emerging market capital flows: The role of global risk and uncertainty (IMF Working Paper WP/20/35). Federal Reserve Bank of Kansas City. [Google Scholar]
  5. Badinger, H., & Reuter, W. H. (2017). The case for fiscal rules. Economic Modelling, 60(1), 334–343. [Google Scholar] [CrossRef]
  6. Bai, J., & Perron, P. (1998). Estimating and testing linear models with multiple structural changes. Econometrica, 66(1), 47–78. [Google Scholar] [CrossRef]
  7. Baker, H. K., Filbeck, G., & Kiymaz, H. (Eds.). (2015). Mutual funds and exchange-traded funds: Building blocks to wealth. Oxford University Press. [Google Scholar]
  8. Baker, M., Bradley, B., & Wurgler, J. (2011). Benchmarks as limits to arbitrage: Understanding the low-volatility anomaly. Financial Analysts Journal, 67(1), 40–54. [Google Scholar] [CrossRef]
  9. Barber, 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(6), 2095–2120. [Google Scholar] [CrossRef]
  10. Bekaert, G., & Wang, X. (2010). Inflation risk and the inflation risk premium. Economic Policy, 25(64), 755–806. [Google Scholar] [CrossRef]
  11. Ben-David, I., Franzoni, F. A., & Moussawi, R. (2014). Do ETFs increase volatility? SSRN Scholarly Paper ID 1967599. Social Science Research Network. [Google Scholar]
  12. Blitz, D., Van Vliet, P., & Baltussen, G. (2019). The volatility effect revisited. The Journal of Portfolio Management, 46(2), 102–121. [Google Scholar] [CrossRef]
  13. Blitz, D., & Vidojevic, M. (2019). The performance of exchange-traded funds. Available online: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3458275 (accessed on 1 April 2025).
  14. Bouri, E., Jain, A., Roubaud, D., & Kristoufek, L. (2021). Does geopolitical risk matter for exchange-traded fund flows in emerging markets? Finance Research Letters, 40, 101703. [Google Scholar] [CrossRef]
  15. Box, T., Davis, R. L., & Fuller, K. P. (2019). ETF competition and market quality. Financial Management, 48(3), 873–916. [Google Scholar] [CrossRef]
  16. Brown, D. C., Davies, S. W., & Ringgenberg, M. C. (2021). ETF arbitrage, non-fundamental demand, and return predictability. Review of Finance, 25(4), 937–972. [Google Scholar] [CrossRef]
  17. Charteris, A. (2013). The price efficiency of South African exchange traded funds. Investment Analysts Journal, 42(78), 1–11. [Google Scholar] [CrossRef]
  18. Chen, N. F., Roll, R., & Ross, S. A. (1986). Economic forces and the stock market. Journal of Business, 5(4), 383–403. [Google Scholar] [CrossRef]
  19. Chiu, J., Chung, H., Ho, K. Y., & Wang, G. H. (2012). Funding liquidity and equity liquidity in the subprime crisis period: Evidence from the ETF market. Journal of Banking & Finance, 36(9), 2660–2671. [Google Scholar] [CrossRef]
  20. Clifford, C. P., Fulkerson, J. A., & Jordan, B. D. (2014). What drives ETF flows? Financial Review, 49(3), 619–642. [Google Scholar] [CrossRef]
  21. Converse, N., Levy-Yeyati, E., & Williams, T. (2023). How ETFs amplify the global financial cycle in emerging markets. The Review of Financial Studies, 36(9), 3423–3462. [Google Scholar] [CrossRef]
  22. Davies, W. (2013). When is a market not a market?: ‘Exemption’, ‘externality’ and ‘exception’ in the case of European state aid rules. Theory, Culture & Society, 30(2), 32–59. [Google Scholar] [CrossRef]
  23. Dlamini, C. S. (2017). The relationship between macroeconomic indicators and stock returns: Evidence from the JSE sectoral indices [Master’s dissertation, University of the Witwatersrand]. [Google Scholar]
  24. Elton, E. J., Gruber, M. J., & De Souza, A. (2019). Passive mutual funds and ETFs: Performance and comparison. Journal of Banking & Finance, 106, 265–275. [Google Scholar] [CrossRef]
  25. Fama, E. F., & French, K. R. (1989). Business conditions and expected returns on stocks and bonds. Journal of Financial Economics, 25(1), 23–49. [Google Scholar] [CrossRef]
  26. Frazzini, A., & Pedersen, L. H. (2014). Betting against beta. Journal of Financial Economics, 111(1), 1–25. [Google Scholar] [CrossRef]
  27. Fulkerson, J. A., Jordan, S. D., & Riley, T. B. (2014). Predictability in bond ETF returns. The Journal of Fixed Income, 23(3), 50. [Google Scholar] [CrossRef]
  28. Fulkerson, J. A., Jordan, S. D., & Travis, D. H. (2017). Bond ETF arbitrage strategies and daily cash flow. The Journal of Fixed Income, 27(1), 49. [Google Scholar] [CrossRef]
  29. Gastineau, G. L. (2004). The benchmark index ETF performance problem. The Journal of Portfolio Management, 30(2), 96–103. [Google Scholar] [CrossRef]
  30. Hamilton, J. D. (1989). A new approach to the economic analysis of nonstationary time series and the business cycle. Econometrica, 57(2), 357–384. [Google Scholar] [CrossRef]
  31. Hoang, K. K. (2018). Risky investments: How local and foreign investors finesse corruption-rife emerging markets. American Sociological Review, 83(4), 657–685. [Google Scholar] [CrossRef]
  32. Hopman, C. (2007). Do supply and demand drive stock prices? Quantitative Finance, 7(1), 37–53. [Google Scholar] [CrossRef]
  33. Höfler, P., Schlag, C., & Schmeling, M. (2023). Passive investing and market quality. Available online: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4567751 (accessed on 21 May 2025).
  34. Islamiah, N. (2025). Macroeconomics and social inequality: Examining the role of purchasing power in affecting income and poverty levels. Golden Ratio of Data in Summary, 5(1), 38–45. [Google Scholar] [CrossRef]
  35. Jame, R., & Tong, Q. (2014). Industry-based style investing. Journal of Financial Markets, 19, 110–130. [Google Scholar] [CrossRef]
  36. Johannesburg Stock Exchange. (2022). Remarks by Lesetja Kganyago, Governor of the South African reserve bank, at the JSE/NYSE market close event, New York. Available online: https://furtherafrica.com/2022/10/14/jse-nyse-to-collaborate-on-dual-listings-etfs-and-digital-assets/ (accessed on 5 December 2024).
  37. Kalaycıoğlu, S. (2004). Exchange traded fund flows. Available online: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=881108 (accessed on 21 May 2025).
  38. Kittur, C. S., Madhhivanan, S., Kumar, K. R., & Sivakumar, B. N. (2025). Do the Macroeconomic variables influence Exchange Traded Funds (ETFs)? Journal of Informatics Education and Research, 5(1), 34–48. [Google Scholar] [CrossRef]
  39. Koepke, R. (2018). Fed policy expectations and portfolio flows to emerging markets. Journal of International Financial Markets, Institutions and Money, 55, 170–194. [Google Scholar] [CrossRef]
  40. Koop, G., & Korobilis, D. (2013). Large time-varying parameter VARs. Journal of Econometrics, 177(2), 185–198. [Google Scholar] [CrossRef]
  41. Krause, T., Ehsani, S., & Lien, D. (2014). Exchange-traded funds, liquidity and volatility. Applied Financial Economics, 24(24), 1617–1630. [Google Scholar] [CrossRef]
  42. Kunjal, D. (2022). The effect of disaggregated country risk on the quality of the South African exchange traded fund market [Unpublished Ph.D. dissertation, University of KwaZulu-Natal]. [Google Scholar]
  43. Kunjal, D., Peerbhai, F., & Muzindutsi, P. F. (2021). The performance of South African exchange traded funds under changing market conditions. Journal of Asset Management, 22(5), 350–359. [Google Scholar] [CrossRef]
  44. Kyle, A. S. (1985). Continuous auctions and insider trading. Econometrica, 53, 1315–1335. [Google Scholar] [CrossRef]
  45. Laopodis, N. T. (2010). Dynamic linkages between monetary policy and the stock market. Review of Quantitative Finance and Accounting, 35(3), 271–293. [Google Scholar] [CrossRef]
  46. Laubscher, P. (2020). A history of South Africa’s modern business cycle. In W. Boshoff (Ed.), Business cycles and structural change in South Africa (pp. 111–166). Springer. [Google Scholar]
  47. Lee, C. M., & Swaminathan, B. (2000). Price momentum and trading volume. The Journal of Finance, 55(5), 2017–2069. [Google Scholar] [CrossRef]
  48. Lerner, P. (2022). The market drives ETFs or ETFs the market: Causality without Granger. arXiv, arXiv:2204.03760. [Google Scholar] [CrossRef]
  49. Lesmeister, S., Limbach, P., Rau, P. R., & Sonnenburg, F. (2022). Indexing and the performance-flow relation of actively managed mutual funds. University of Cologne, Centre for Financial Research (CFR). [Google Scholar]
  50. Levy, B., Hirsch, A., Naidoo, V., & Nxele, M. (2021). South Africa: When strong institutions and massive inequalities collide. Carnegie. [Google Scholar]
  51. Lo, A. W. (2004). The adaptive market hypothesis. Journal of Portfolio Management, 30(5), 15–29. [Google Scholar] [CrossRef]
  52. Lo, A. W., & MacKinlay, A. C. (1988). Stock market prices do not follow random walks: Evidence from a simple specification test. The Review of Financial Studies, 1(1), 41–66. [Google Scholar] [CrossRef]
  53. Lubik, T. A., & Matthes, C. (2015). Time-varying parameter vector autoregressions: Specification, estimation, and an application. Federal Reserve Bank of Richmond Economic Quarterly, 101(4), 323–352. [Google Scholar] [CrossRef]
  54. Mofokeng, T., & Moodley, F. (2025). The effects of changing macroeconomic conditions on the commercial bank returns in South Africa. International Journal of Economics and Financial Issues, 15(2), 76–86. [Google Scholar] [CrossRef]
  55. Moodley, F., Ferreira-Schenk, S., & Matlhaku, K. (2024). Effect of market-wide investor sentiment on South African Government bond indices of varying maturities under changing market conditions. Economies, 12(10), 265. [Google Scholar] [CrossRef]
  56. Moodley, F., Nzimande, N., & Muzindutsi, P. F. (2022). Stock returns indices and changing macroeconomic conditions: Evidence from the Johannesburg securities exchange. The Journal of Accounting and Management, 12(3), 147–161. [Google Scholar]
  57. Moskowitz, T. J., & Grinblatt, M. (1999). Do industries explain momentum? The Journal of Finance, 54(4), 1249–1290. [Google Scholar] [CrossRef]
  58. Nadler, D., & Schmidt, A. B. (2015). Impact of macroeconomic announcements on ETF trading volumes. Journal of Trading, 10, 31–35. [Google Scholar] [CrossRef]
  59. Nakajima, J. (2023). The impact of macroeconomic uncertainty on the relationship between financial volatility and real economic activity. Applied Economics, 56, 5591–5604. [Google Scholar] [CrossRef]
  60. Narend, S. (2014). Performance of ETFs and index funds: A comparative analysis (Vol. 1, p. 17). National Stock Exchange of India Ltd. [Google Scholar]
  61. Nkomo, T., & Moodley, F. (2025). Do geopolitical risk and market conditions drive JSE sector returns? Management and Economics Review, 10(2), 324–341. [Google Scholar] [CrossRef]
  62. Peltomäki, J. (2017). Beta as a determinant of investor activity in sector exchange-traded funds. The Quarterly Review of Economics and Finance, 65, 137–145. [Google Scholar] [CrossRef]
  63. Rapach, D. E., Strauss, J. K., & Zhou, G. (2013). International stock return predictability: What is the role of the United States? The Journal of Finance, 68(4), 1633–1662. [Google Scholar] [CrossRef]
  64. Rompotis, G. G. (2011). Predictable patterns in ETFs’ return and tracking error. Studies in Economics and Finance, 28(1), 14–35. [Google Scholar] [CrossRef]
  65. Shenjere, P. A., Ferreira-Schenk, S., & Moodley, F. (2025). Does investor sentiment influence South African ETF flows during different market conditions? Economies, 13(1), 10. [Google Scholar] [CrossRef]
  66. South African Reserve Bank. (2020). Quarterly bulletin—Financial markets developments. SARB. [Google Scholar]
  67. World Bank. (2020). South Africa economic update: From containment to recovery. World Bank. [Google Scholar]
  68. Wu, C., Xiong, X., & Gao, Y. (2021). Performance comparisons between ETFs and traditional index funds: Evidence from China. Finance Research Letters, 40, 101740. [Google Scholar] [CrossRef]
Figure 1. Smooth Transition Probabilities. Source: Authors’ own estimation (2025).
Figure 1. Smooth Transition Probabilities. Source: Authors’ own estimation (2025).
Economies 13 00260 g001
Table 1. Sample of ETF and Macroeconomic variables.
Table 1. Sample of ETF and Macroeconomic variables.
VariableAbbreviationDescriptionStart Date
Panel A: ETF Returns
Invest Top 40 ETFETFT40Tracks the JSE Top 40 index by market capitalization10/2010
Invest SWIX 40 ETFETFSWXTracks the JSE All-share index10/2010
Satrix SWIX Top 40 ETFSTX40Tracks the JSE shareholders-weighted Top 40 index11/2000
Sygnia Itrix Top 40 ETFSTXSWXTracks the JSE Top 40 index by market return 04/2006
Panel B: Macroeconomic Variables
Inflation rate C P I -
Money supply (M2) M 2 -
Short-term interest S T I N T -
Long-term interest L T I N T -
Gross domestic product G D P -
Real effective exchange rate R E E R -
Notes: Authors’ own compilation (2025).
Table 2. Descriptive Statistics and Preliminary results.
Table 2. Descriptive Statistics and Preliminary results.
ETFSWXETFT40STX40STXSWXCPIM2ST_INTLT_INTGDPREER
Panel A: Descriptive Statistics
Mean0.0052080.0061210.0060890.0052860.2272920.5193380.2927760.2573330.513162−0.101130
Median0.0079100.0080450.0074790.0062970.1848660.3647860.190949−0.0559300.556007−0.421891
Maximum0.1470440.1343210.1302070.1375031.4541396.28622113.7301317.7352024.353327.171398
Minimum−0.134555−0.110446−0.118299−0.142629−1.753214−3.108036−24.20321−10.12012−20.06378−6.451613
Std. Dev.0.0435010.0420120.0405960.0403210.4709161.4456213.8497703.2700992.8029032.862607
Skewness−0.1268120.1076170.191046−0.079885−0.4545320.409115−1.4759021.2558151.3781760.098581
Kurtosis4.0050533.2941023.6348384.0636755.9552773.78611713.951199.44500552.569262.532677
Jarque–Bera7.0735060.8744093.6143397.61647062.937068.475909846.8911314.988616226.001.693655
Probability0.0291080.6458390.1641180.0221870.0000000.0144370.0000000.0000000.0000000.428773
Observations158158158158158158158158158158
Panel B: Variance Inflation Factor Test
VIF Stat----1.0614241.1036701.0909841.1524341.1122971.177853
Panel C: Unit Root and Stationarity Tests
ADF−14.95804 ***−14.47077 ***−14.15896 ***−14.30402 ***−6.594083 ***−14.065538 ***−5.935803 ***−10.79380 ***−8.114136 ***−10.28854 ***
KPSS0.2236560.1020270.1042960.2317540.2928260.3109340.1537460.0590440.1502030.117962
ADF Break−16.07975 ***−15.21584 ***−15.12700 ***−15.57749 ***−8.814163 ***−13.66101 ***−10.47896 ***−11.92506 ***−12.77315 ***−11.20548 ***
Notes: 1. *** indicate a statistic significance level of 1%. 2. Source: Authors own estimation (2025).
Table 3. Unconditional Correlation.
Table 3. Unconditional Correlation.
VariablesETFSWXETFT40STX40STXSWX
CPI−0.021282 *−0.083164−0.066460 ***−0.028252 **
(0.0907)(0.2989)(0.0067)(0.0246)
M2−0.077552−0.021709 ***−0.067581 ***−0.055524 ***
(0.3328)(0.0066)(0.0088)(0.0084)
ST_INT−0.033329 *−0.034433 *−0.037689−0.041607 ***
(0.0776)(0.0676)(0.6382)(0.0037)
LT_INT−0.281338 ***−0.231347 ***−0.281262 ***−0.265631 ***
(0.0003)(0.0034)(0.0003)(0.0007)
GDP−0.031104 *−0.052772 **−0.053672−0.061432 **
(0.0980)(0.0102)(0.5030)(0.0432)
REER−0.091519 *−0.130053−0.127628 **−0.114478
(0.0528)(0.1034)(0.0100)(0.1521)
Notes: 1. The parenthesis provides the associated p-values. 2. *, **, and *** indicate a statistic significance level of 10%, 5%, and 1%, respectively. 3. Source: Authors’ own estimation (2025).
Table 4. Markov Regime-Switching Model Results.
Table 4. Markov Regime-Switching Model Results.
ETFSWXETFT40STX40STXSWX
Panel A: Bull Regime
C0.036534 *0.010367 ***0.008060 **0.002519
CPI0.063551 *0.028374 ***−0.002166 *0.008454
M2−0.013067−0.059574 ***−0.001047−0.002923
ST_INT0.017883 *0.030833 ***0.0002780.002746 ***
LT_INT−0.006992 ***−0.030854 ***−0.004717 ***−0.003955 ***
GDP0.0022300.002190 ***−0.0025690.002473 *
REER−0.004261 ***−0.007413 ***−0.002981 **−0.002808 **
σ −3.178687 ***−1.493026 ***−3.258975 ***−3.391161 ***
Panel B: Bear Regime
C−0.028047 *−0.007904 *0.014071 ***0.026780 ***
CPI−0.050439−0.0125270.012067 ***−0.068065 ***
M20.001128 **0.0030230.005385 ***0.030410 ***
ST_INT−0.001890−0.0002360.001108 ***−0.005635 ***
LT_INT0.006153−0.003118 **−0.003297 ***−0.007997 ***
GDP−0.001852 **−0.0019350.001110 ***0.005509 ***
REER0.002242−0.001948−0.005293 ***−0.003836 ***
σ −3.772294 ***−6.203929 ***−8.571630 ***−4.932839 ***
Panel C: Diagnostic Tests
DW-STAT2.3182322.3052092.2199542.308811
LM-Stat1.3526843.4668464.0177313.786891
p Value0.12530.16500.13410.1506
Notes: 1. *, **, and *** indicate a statistic significance level of 10%, 5%, and 1%, respectively. 2. Source: Authors’ own estimation (2025).
Table 5. Transition Probabilities and Expected Duration Results.
Table 5. Transition Probabilities and Expected Duration Results.
ETFSWXETFT40STX40STXSWX
Panel A: Bull Regime
P110.9504370.1360460.9331640.810897
T1120.176421.15746914.962055.288122
Panel B: Bear Regime
P220.7669830.8572610.1560159.30 × 10−9
T224.2915267.0058161.1848551.000000
Notes: 1. P11 and P22 reflect the constant transition probabilities associated with a bull and bear regime, respectively. 2. T11 and T22 provide the expected duration in a bull and bear regime, respectively. 3. Source: Authors’ own estimation (2025).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Moodley, F.; Lawrence, B.; Tabash, M.I. Are Macroeconomic Variables a Determinant of ETF Flow in South Africa Under Different Economic Conditions? Economies 2025, 13, 260. https://doi.org/10.3390/economies13090260

AMA Style

Moodley F, Lawrence B, Tabash MI. Are Macroeconomic Variables a Determinant of ETF Flow in South Africa Under Different Economic Conditions? Economies. 2025; 13(9):260. https://doi.org/10.3390/economies13090260

Chicago/Turabian Style

Moodley, Fabian, Babatunde Lawrence, and Mosab I. Tabash. 2025. "Are Macroeconomic Variables a Determinant of ETF Flow in South Africa Under Different Economic Conditions?" Economies 13, no. 9: 260. https://doi.org/10.3390/economies13090260

APA Style

Moodley, F., Lawrence, B., & Tabash, M. I. (2025). Are Macroeconomic Variables a Determinant of ETF Flow in South Africa Under Different Economic Conditions? Economies, 13(9), 260. https://doi.org/10.3390/economies13090260

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop