Journal of Islamic Monetary Economics and Finance.
Vol.
No.
, pp.
535 - 582
p-ISSN: 2460-6146, e-ISSN: 2460-6618
DIVERSIFYING ISLAMIC HAVEN ASSETS
Bayu Adi NugrohoO Independent Researcher.
Indonesia
ABSTRACT
This study reassesses the safe haven properties of gold and Sukuk using a new framework that incorporates nonstationary volatility and proposes a trading strategy to construct a gold Ae Sukuk Ae Islamic equities portfolio that can outperform the hardto-beat nayve method and the covariance-based approaches.
In line with previous studies, it employs data of four exchange-traded funds: Dow Jones Global Sukuk.
Wahed FTSE USA Shariah.
MSCI Emerging Market Islamic, and SPDR Gold.
In the study, an enhanced version of wavelet quintile correlation is proposed to re-evaluate the haven qualities of gold and Sukuk.
The results show that gold and Sukuk are safe haven assets.
Next, applying a dual momentum strategy, we demonstrate that the riskadjusted returns of our proposed trading strategy outshine the nayve method and the covariance-based approaches.
Our research employs real returns and a rolling window approach to avoid money illusion, overfitting, look-ahead bias, and flawless hindsight.
The main results prevail in the robustness tests.
Keywords: Dual momentum.
Quantile correlation.
Wavelet.
JEL classification: G11.
G12.
G14.
Article history:
Received : August 20, 2024 Revised : December 4, 2024 Accepted : August 29, 2025 Available online : September 30, 2025 https://doi.
org/10.
21098/jimf.
O CONTACT Bayu Adi Nugroho: bayunugrohomito@gmail.
Independent Researcher.
Indonesia Diversifying Islamic Haven Assets INTRODUCTION Previous studies have examined the quality of gold and bonds as haven assets.
While the view that gold and bonds are haven assets has been well supported by numerous studies, adverse shocks as severe as the COVID-19 pandemic and the Russia-Ukraine war have placed this view into doubt .
In this spirit, this research revisits the hypothesis of gold and Islamic bonds or Sukuk as haven assets using a new framework and trading strategy.
Kumar & Padakandla .
develop a wavelet quintile correlation that can show the haven quality of gold and Sukuk on different trading days and quintiles.
However, the method relies on a single regime volatility.
Financial assets contain structural breaks that contribute to variations in volatility, and disregarding this feature can significantly reduce the accuracy of volatility estimates (Ardia et al.
, 2.
Hence, this studyAos first and primary objective is to re-evaluate the haven quality of gold and Sukuk using an enhanced version of a wavelet quintile correlation approach.
Once we know the haven quality of gold and Sukuk on different trading days and quintiles, the next discussion is how to allocate the haven assets into a portfolio of Islamic equities.
This is also important since many ways exist to construct an optimal gold Ae Sukuk Ae Islamic equities portfolio.
Therefore, the second objective is to propose a trading strategy that can outperform a hard-to-beat nayve strategy and covariance-based methods widely used in the literature.
The current paper extends the study of Kumar & Padakandla .
to reexamine the role of gold and Sukuk in Islamic equities portfolio, whether they are haven assets.
However, there are two important differences between this research and that of Kumar & Padakandla .
First, this research uses real returns.
The objective of using inflation-adjusted returns is to avoid money illusion.
Second, the current research examines the role of gold and Sukuk in Islamic equities portfolios, taking into consideration of nonstationary volatility.
In addition to the above two, the current study is also interested in understanding the composition of gold Ae Sukuk Ae Islamic equities in a portfolio, and this is not a trivial issue.
Hence, the current research should also be seen as extending the work of Vliet & Lohre .
to gold Ae Sukuk Ae Islamic equities portfolio.
However, this paper uses dynamic weights instead of the constant weight strategy used by Vliet & Lohre .
To my knowledge, the introduction of nonstationary volatility on wavelet quantile correlation and dynamic weight approach on gold Ae Sukuk Ae Islamic equities portfolio has not been explored.
The new approaches offer several fresh insights.
First, gold and Sukuk are haven assets.
Second, the evidence of a safe-haven status is also associated with portfolio allocation.
a portfolioAos combination of gold Ae Sukuk Ae Islamic equities produces significantly lower risk than all Islamic equities portfolios.
Finally, a dynamic weight strategy using a dual momentum approach is better than a constant weight strategy.
When volatility is very high, stopping loss by converting assets to cash is necessary.
The articleAos structure is as follows: The literature review begins with the definition of a haven asset and then discusses the empirical research on the abilities of gold and Sukuk to reduce risk.
The third section is the methodology, followed by results and analysis.
Lastly.
I conclude and recommend future research and Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 II.
LITERATURE REVIEW
This section reviews the literature on gold and SukukAos ability to reduce equities It also deliberates the limitations of previous research that this research Baur & Lucey .
apply a quantile regression approach and find that gold is a haven asset for equities and non-Islamic bonds in the US.
UK, and Germany, both in normal and extreme market conditions.
Likewise, applying a similar approach.
Baur & McDermott .
find that gold is a haven asset for most equities in the developed markets, while gold is not a haven for equities in emerging markets.
In addition.
Robiyanto et al.
implement a quantile regression approach and conclude that gold is a haven asset for investors concerned about ethics in Indonesia.
In the literature.
GARCH-based models are extensively used to evaluate the haven properties of gold.
Izadi & Hassan .
find that gold futures minimize the risk of the equities portfolio.
Similarly.
Raza et al.
showed that gold is more effective in hedging Islamic equities than non-Islamic equities.
Additionally.
Akhtaruzzaman et al.
indicate that gold is a haven asset at the beginning phases of COVID-19.
Further.
Salisu et al.
indicate that adding gold to an equities portfolio improves the risk-adjusted returns.
Bahloul et al.
indicated that gold is a strong haven in developed markets.
Only a few papers conclude that gold is not a haven asset.
For example.
Bandhu Majumder .
concludes that gold is not a haven asset for Indian Furthermore.
Corbet et al.
indicate that gold does not function as a instead, it acts as a booster of contagion.
In addition.
Ghaemi Asl & Rashidi .
employ VAR-BEKK-GARCH to examine the safe-haven features of Sukuk in Middle Eastern and North African countries (MENA).
They find that Sukuk is a haven asset.
Similarly.
Shahzad et .
, utilizing a regime-switching copula technique, find that Sukuk is an excellent haven asset.
They propose a portfolio comprising 50% Sukuk and 50% Islamic shares.
When there is increased uncertainty, risk-averse investors shift their investment to Sukuk.
The literature on constructing a goldAiSukuk Ae Islamic equities portfolio is non-existent.
The nearest study is Vliet & Lohre .
They construct goldAi bonds Ai and non-Islamic equities portfolios and argue that gold is a volatile Hence, the optimal weight of gold in the portfolio is around 10 %, which is determined based on from a constant-weight strategy.
The summary of the selected literature is in Table 1.
The table shows that previous researchers use various econometric models to evaluate the safe-haven status of Gold.
Sukuk, and Islamic equities: a vine-copula framework, dynamic GARCH families, wavelet analysis.
Markov-switching copula, wavelet-based quintile, and quantile regression.
Kumar & Padakandla .
recently propose a new approach, namely theWavelet Quantile Correlation or WQC, to cope with the weaknesses of the previously stated models.
Simply put, the WQC can identify safe-haven characteristics over various trading timeframes.
However, the WQC proposed by Kumar & Padakandla .
does not consider nonstationary Diversifying Islamic Haven Assets Therefore, there is a weakness in the original WQC approach.
The volatility pattern of financial returns can vary due to shocks such as COVID-19 and the Russia Ae Ukraine conflict.
The conventional GARCH models however mostly assume stationary.
Therefore, this research evaluates gold and SukukAos safeAe haven properties based on nonstationary volatility.
Two relatively new methods to assess the volatility pattern of financial returns are the MarkovSwitching GARCH or MSGARCH and the Time-Varying GARCH or TVAeGARCH (Ardia.
Bluteau.
Boudt, et al.
, 2019.
Campos-Martins & Sucarrat, 2.
Our research adds the development of the methods used to evaluate the haven properties of gold and Sukuk, offering a new framework.
Further, the advantages of gold Ae Sukuk Ae Islamic equities portfolio are still unknown.
Thus, addressing these issues is the gap that this research fills.
Table 1.
A Selected Review of the Literature Author.
Methods, period of study, and sample Result Limitations of the study Gold as a haven asset Baur & Lucey Quintile regression.
Period Gold is a haven asset for Notably, other researchers .
of the study: 1995 to 2005.
equities and non-Islamic have highly cited this Sample: the US, the UK.
Bonds in the US.
UK, and research concerning the German equities, and non- Germany, both on normal definition of safe-haven.
Islamic Bonds.
and during extreme market hedge, and diversifier.
However, the authors do not account for regimeswitching volatility or inflation without a trading Bandhu Various Vector Auto Gold is not a haven asset The author do not account Regression (VAR) Ae BEKK for the Indian equities.
for inflation.
Majumder .
Ae GARH models.
Period of the study: Dec 2010 to Dec Sample: Indian stock Izadi & Hassan Dynamic conditional Gold futures minimize The authors do not account the risk of the equities for regime-switching .
correlation Ae GARCH.
Period of the study: Jan volatility or inflation and were without a trading 2000 to Oct 2014.
Sample:
commodity and equity markets of G7 nations.
Raza et al.
DCC.
ADCC, and GOGold is more effective in The authors do not account GARCH.
Period of the hedging Islamic equities for regime-switching study: 1996Ae2015.
Sample: than non-Islamic equities.
volatility or inflation and were without a trading Islamic and non-Islamic equities from Dow Jones.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Table 1.
A Selected Review of the Literature (Continue.
Author.
Nugroho .
Methods, period of study, and sample A combination of Markovswitching GARCH Ae Copula.
Period of the study: Dec 2015 to June Sample: Islamic and non-Islamic indices.
Result Limitations of the study Gold improves the Although the author value-at-risk of gold - the applies a MarkovIslamic equities portfolio switching GARCH more than the non-Islamic to account for regime equities portfolio.
changes, the author utilizes a nominal return without a trading strategy.
Corbet et al.
DCC Ae GARCH.
Period Gold does not function as a The authors use nominal .
of the study: March 2019 instead, it acts as a returns and do not account to March 2020.
Sample:
booster of contagion.
for regime-switching volatility, which is very hourly data of ChinaAos equity market, gold, and persistent in the early phase of COVID-19.
Bitcoin.
Adding gold to an equities The hedge ratios involved Salisu.
Vo, & Optimal weights and Lucey .
hedge ratios based on portfolio improves the risk- shortAeselling.
Conventional adjusted returns.
shortAeselling is not allowed VARMA Ae GARCH.
Period by Islamic standards.
The of the study: Jan 2019 to Jul 2020.
Sample: the US authors also use nominal returns and assume no sectoral indices and gold.
transaction costs for the trading strategy.
The authors use nominal Baur & Quintile regression.
Period Gold is a haven asset for McDermott of the study: 1995 to 2005.
most equities in developed returns and make no .
Sample: non-Islamic markets but not for equities recommendation regarding in emerging nations.
the percentage of gold in a equities of G7 nations.
BRICS.
Australia, and Switzerland.
Akhtaruzzaman Optimal weights and Gold is a haven asset at the The recommended et al.
hedge ratios based on very beginning phases of optimal weight of gold DCC Ae GARCH and COVID-19.
in the portfolio is 56%.
the quintile regression.
Still, it is not economical Period of the study: Dec for investors to rebalance 2019 to Apr 2020.
Sample:
their portfolio hourly.
The hourly returns of gold and authors also use nominal conventional equity indices returns and assume no of the USA.
Europe.
Japan, transaction costs.
and China.
Baur & Smales Linear and GARCH-based Gold acts as a consistent The authors use the signal Period of the haven asset against of the geopolitical risk study: Jan 1985 to Oct geopolitical risk.
index as a trading strategy Sample: precious without considering metals futures and S&P 500 transaction costs.
The daily and monthly signals may generate high transaction Diversifying Islamic Haven Assets Table 1.
A Selected Review of the Literature (Continue.
Author.
Salisu.
Vo, & Lawal .
Methods, period of study, and sample Optimal portfolio weights based on VARMAAe GARCH.
Period of the study: Jan 2016 to Aug Sample: precious metals and crude oil.
Result Gold serves as a haven asset towards the oil price Limitations of the study The optimal portfolio weight strategy is outdated and impractical since it involves daily rebalancing .
mplying high transaction The authors also use nominal returns.
The authors use nominal Gold serves as a strong Bahloul et al.
DCC-GARCH-based returns, do not account for haven in developed .
regression of Ratner & regime-switching volatility.
Interestingly.
Chiu .
Period of Islamic equities act as weak offer no trading strategy, the study: May 2015 to and assume no transaction May 2020.
Sample: Gold, haven assets against the Simply put, the Islamic, and non-Islamic risk of non-Islamic equities equities from the US.
Italy, in Spain.
France, the UK, authors do not recommend and Malaysia.
the proportion of gold in a Russia.
France, the UK.
Malaysia.
Spain.
China.
Germany, and Brazil.
The inclusion of gold in This research only focuses Vliet & Lohre Constant weights .
% .
Gold, 45% conventional a typical bondAeequities on the US market with portfolio reduces risk.
lower inflation rates than bonds, 45% low-volatility equitie.
with yearly emerging markets in 2020 - 2022.
Period of the study: 1975 to 2022.
Sample: bonds and equities in the US.
Kumar & A newly developed Gold serves as a haven The authors use nominal Padakandla avelet quintile returns and do not account Period of the for regime-switching study: Jan 2015 to Dec Sample: non-Islamic equities from developed markets (France, the US.
India, and Europ.
, gold, and Bitcoin.
Rusmita et al.
A threshold GARCH Gold is a haven asset.
The authors use nominal .
approach (TGARCH) and a returns and do not account quintile regression.
Period for regime-switching of the study: Jan 2011 to Oct 2022.
Sample: Antam gold and the Jakarta Islamic Index.
Sukuk as a diversifier Nugroho & A novel hedge ratio Sukuk acts as a diversifier.
The authors use nominal Kusumawardhani involving a modified returns and do not account EWMA approach.
DECOfor regime-switching GARCH, and wavelet.
volatility, making the Period of the study: Jan trading strategy from 2020 to Oct 2023.
Sample:
the hedge ratio approach Islamic exchange-traded Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Table 1.
A Selected Review of the Literature (Continue.
Author.
Naeem et al.
Qadri et al.
Methods, period of study, and sample Result Limitations of the study Asymmetric DCC-GARCH Sukuk acts as a diversifier.
The limitations of this regression of Ratner & study are similar to Chiu .
and hedge those in Nugroho & Period of the study:
Kusumawardhani .
Jan 2020 to Oct 2023.
Sample: Dow Jones Sukuk Bond Index, green bonds, and equities from ten GARCH-based regression.
Sukuk is not a haven asset.
The authors use nominal returns and do not account Period of the study: Aug 2012 to Jun 2022.
Sample:
for regime-switching sukuk and conventional bond indices from various i.
METHODOLOGY
Data This study uses the following exchange-traded funds (ETF.
: Dow Jones Global Sukuk.
FTSE USA Shariah, iShares MSCI EM Islamic, and SPDR Gold.
The underlying assets for Sukuk ETF include KSA Sukuk Limited, issued by Saudi Arabia, with a profit rate of 3.
The average maturity of the Islamic bonds in the ETF is about eight years.
The daily data covers the sample from Jan 02, 2020, to Dec 29, 2023.
This sample period is dictated by data availability.
The data are sourced from https://w.
com/, which is a reliable data source and used by others (Liu et al.
, 2.
The returns are calculated by taking a natural log of todayAos price divided by yesterdayAos price.
In the literature, few studies adjust nominal returns with inflation.
Along this line, this research uses real returns.
The inflation data for the USA market are from org while those for the emerging markets are from w.
All prices are in US dollars to eliminate bias from foreign exchange fluctuations.
This research utilizes those ETFs for the following reasons: First, previous studies concluded that Sukuk are not safe-haven assets .
ee Table .
However, prior results do not consider dual-regime volatility.
Hence, this study compares the performance of a previous model and our model using the same data set.
Second.
ETFs can offer lower operating costs than traditional open-end funds, flexible trading, and greater transparency.
Nonstationarity Volatility Models Ae Wavelet Quintile Correlation (MSGARCH-WQC and TVGARCH-WQC) The conventional WQC does not account for nonstationary volatility.
Financial assets contain structural breaks that contribute to their volatility movement, and disregarding this feature can significantly reduce the accuracy of volatility Diversifying Islamic Haven Assets estimates (Ardia et al.
, 2.
TV-GARCH is more robust when modeling nonstationary volatility (Campos-Martins & Sucarrat, 2.
Hence, this study proposes MSGARCH-WQC and TVGARCH-WQC.
1 Markov-switching GARCH (MSGARCH) This study follows a two-step methodology to estimate MSGARCH-WQC.
First, we utilize MSGARCH and an AR .
filter to exclude autoregressive influences from the data before estimating the models based on the residuals (Ardia.
Bluteau, & Ryede, 2.
The improved risk estimates of the model stem from their ability to adjust to changes in the unconditional volatility level swiftly.
MSGARCH
can be estimated using the Maximum Likelihood method.
Nonetheless, several current studies show certain benefits to using a Bayesian approach (Ardia.
Kolly, et al.
, 2017.
Casarin et al.
, 2.
For instance, the Bayesian approach allows investigatation of the joint posterior distribution of the model parameters.
Hence, we apply a Bayesian technique to estimate the model parameters using Markov chain Monte Carlo (MCMC) simulations.
In addition, this study incorporates E-GARCH for the scedastic specification (Nelson, 1.
We use dual regimes ( = .
Compared to symmetric GARCH, the E-GARCH, which is an asymmetric GARCH, fits data better.
Haas.
Mittnik, & Paolella .
incorporate the scedastic specification into the MSGARCH Still, the model is more robust by permitting innovations to come from distributions other than normal (Cerqueti et al.
, 2.
For the innovations, we thus use the Student- and the Skewed Student- .
To save space, the interested reader is referred to Ardia.
Bluteau.
Boudt, et al.
to estimate MSGARCH using the R statistical program.
2 Time-varying GARCH (TVGARCH) This section briefly explains the test of nonstationary GARCH.
Since the MSGARCH is conducted on univariate settings, the TVAeGARCH approach is implemented Verifying whether the unconditional variance is time-invariant is crucial before assessing a TV-GARCH model.
Rejecting the null hypothesis (H.
indicates nonstationarity, suggesting that a conventional GARCH model with constant parameters is not appropriate and consequently not suited to fit the data.
The unconditional variance under the alternative hypothesis is time-varying.
where c0*, c1*, c2*, and c3* are functions of the initial parameters.
is the number of transitions that need to be identified.
Ho holds if c1* = c2* = c3* = 0.
To save space, the interested reader is referred to Campos-Martins & Sucarrat .
to estimate TVGARCH using the R statistical program.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Wavelet Quintile Correlation (WQC) We obtain WQC with the following steps.
First, we receive the Quantile Correlation (QC) of two variables and based on Li et al.
We briefly discuss QC Let be the Eth quantile of and be the Eth quantile of conditional is independent of if and only if the random variables I ( ) > 0 and are independent.
I(.
) is the indicator function.
For 0 < E < 1, the quantile covariance is:
> .
, .
= I( .
The QC is:
The QC shows the correlation between asset pairs across different quantiles.
asset with safe-haven qualities should negatively correlate with another asset in a turbulent market, which QC will identify in the lower quantiles.
Additionally, we assume that investors have different preferences over different time horizons when selecting a safe-haven asset.
We derive these dynamics by looking at the dependence structure over multiple timescales.
As a result, this study utilizes Wavelet Quantile Correlation (WQC).
We use a maximum overlapping discrete wavelet transform (MODWT), as Percival & Walden .
suggest, to decompose the asset pairs.
we briefly discuss MODWT below.
Let [ ] be a signal of length T, such that T=2J for some integer .
An orthogonal wavelet defines the low-pass and high-pass filters, [ ] and [ ], respectively.
produce the approximation coefficients [ ] and [ ], [ ] is convolved with [ ] at the first level and with [ ] at the second.
Next, we apply the same strategy to filter [ ], but we utilize modified filters , which we get by dyadic up-sampling For yeo = 1, 2.
Ae 1, where O yeo, the coefficients are computed as follows:
Diversifying Islamic Haven Assets applying a yeo level decomposition to is the up-sampling operator.
After , the WQC is obtained as follows:
Portfolio Construction Methods Nayve (The Benchmar.
The first method for portfolio optimization used in this study is 1/N or nayve DeMiguel et al.
illustrate that errors in measuring means and covariances undermined all the benefits from optimal diversification instead of naive diversification.
Minimum Variance The calculation of the minimum-variance weight is as follows (Ardia et al.
, 2.
is the long-only investment constraint.
Risk Parity All of the assets in an equal-risk-contribution portfolio contribute equally to the total volatility of the portfolio.
In other words, itAos the portfolio where each assetAos percentage contribution to volatility risk equals 1/N.
The calculation of the portfolio weight is as follows (Maillard et al.
, 2.
Where Dual Momentum Dual momentum assigns a relative value to each asset class based on how well it has performed over the past three months relative to other assets in the same class and whether or not it has had a positive return.
As long as the top-performing asset in the asset class has a positive return above zero, dual momentum invests in those assets.
Otherwise, the allocation is shifted to cash.
This strategy is inspired by Antonacci .
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Rolling Window Approach We run the portfolio construction methods in a rolling window approach to avoid overfitting, look-ahead bias, and ideal hindsight.
Furthermore, we intentionally make our analysis straightforward and non-exclusive.
Our fundamental rationale for taking this method is to verify that the strategies presented are practical for Hence, this research uses the following steps to generate the portfolio weights:
To determine the parameters in the optimization, we use the closest M=250 daysAo data before the rebalancing time t.
Solve the relevant optimization problem to obtain the weights at t.
Following the holding period .
lso known as the rebalancing perio.
, rebalance the weights by carrying out procedures .
IV.
RESULTS AND ANALYSIS
Descriptive Statistics Table 2 presents descriptive statistics of the variables.
This study adjusts nominal returns with inflation.
The table implies that inflation in the emerging nations is higher, indicating that, by the drawdown, an investmentAos peak-to-trough decreases over a given period.
Inflation also affects the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
The table also reveals that the kurtosis values are greater than 3, implying a leptokurtic distribution.
Interestingly.
Sukuk has a positive skewness.
Skewness is the third central moment, commonly used to assess the distributionAos divergence from symmetry.
Assets with a positive skew typically have minor losses and few significant gains.
On the contrary, negatively skewed assets usually have many small gains and few significant losses.
MSGARCH
This study employs dual-regime E-GARCH with the Student- and the Skewed Student- for the innovations.
Our selection is appropriate based on the deviance information criterion (Table .
Table 4 reports the parameters from E-GARCH with the Skewed Student- distribution up to two regimes.
As expected, the unconditional volatility .
or UV of yeo = 2 is higher than yeo = 1.
In particular, the Islamic equities in the emerging markets have higher volatility.
Moreover, the values of in yeo = 2, are higher than in the single-regime .
eo = .
Specifically, the long-term volatility in the USA market is higher than in the emerging markets.
Both regimes are very persistent, with posterior probabilities of A11 and A22 higher than 95%.
Further.
Figure 1 clearly illustrates that the high volatility regime .
ed line.
occurred during the early phase of the COVID-19 pandemic and the RussiaUkraine conflict.
Diversifying Islamic Haven Assets Table 3.
Deviance Information Criterion
Emerging Markets single-regime Standard GARCH
E-GARCH
dual-regime Standard GARCH
E-GARCH
USA
single-regime Standard GARCH
E-GARCH
dual-regime Standard GARCH
E-GARCH
Normal Student- Skewed Student- Notes: The highlighted values show that the dualAeregime MSGARCH model outperforms the singleAeregime model.
Smoothed probabilities - Islamic Equities from Emerging Markets Volatility (%) - Islamic Equities from Emerging Markets Figure 1.
Smoothed Probabilities and Volatility Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Smoothed probabilities - Islamic Equties from USA Volatility (%) - Islamic Equities from USA Figure 1.
Smoothed Probabilities and Volatility (Continue.
USA Emerging USA Emerging real returns Gold Sukuk nominal returns real returns .
merging Gold Sukuk Gold Sukuk real returns (USA) Notes: This table shows the basic statistics of the return series.
Real returns indicate that nominal returns have been adjusted to inflation.
Kurtosis values were greater than 3, implying a leptokurtic distribution.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more substantial the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
The daily data covers the sample from Jan 02, 2020, to Dec 29, 2023.
Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % nominal returns Table 2.
Descriptive Statistics Diversifying Islamic Haven Assets Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Table 4.
Parameter Estimates yeegime .
eo = .
A11
eo = .
single-regime USA Emerging USA dual-regime Emerging Notes: The table shows the mean of the posterior sample for the E-GARCH model and Skewed Student- with single and are the tail and asymmetry values, respectively.
The estimates are and dual regimes.
The parameters of is the unconditional volatility.
E-GARCH estimates in regime yeo are computed as ln taken from 1000 draws.
The values of indicate long-term volatility.
The Conventional Wavelet Quintile Correlation (WQC) If an asset has safe-haven features, its correlation with other assets must be negative during market turmoil, particularly at lower quantiles.
Figure 2 shows that QC is positive across trading periods and lower quintiles for Gold/Islamic equities in emerging markets.
QC is also positive across the median quintiles.
These results imply that gold is not a haven asset but a diversifier for Islamic equities in emerging markets.
Also.
Figure 2 shows that QC is positive across trading periods and lower quintiles for Gold/Islamic equities in the US.
However.
QC is negative at the median quintiles for one trading year.
These results suggest that gold is not a haven asset but a strong hedge for Islamic equities in the US.
Moreover.
Figure 3 shows that QC is positive across trading periods and quintiles for Sukuk/Islamic equities in the US.
Similarly.
QC is positive across trading periods and quintiles for Sukuk/Islamic equities in emerging markets.
These results indicate that Sukuk is a diversifier for Islamic equities in emerging markets and the USA.
Diversifying Islamic Haven Assets MSGARCH Ae WQC After accounting for dual-regime volatility.
Figure 4 illustrates that QC is negative for 32-64 days at the lower quintiles for Gold/Islamic equities in the USA.
QC is also negative across several trading days at the median quintiles.
These results imply that gold is a haven asset and a hedge for Islamic equities in the USA.
Moreover.
Figure 4 also illustrates that QC is negative for 32-64 days at the lower quintiles for Gold/Islamic equities in emerging markets.
However.
QC is positive across the median quintiles.
These results imply that gold is a haven asset and a diversifier for Islamic equities in emerging markets.
Further.
Figure 5 indicates that QC is negative for 2-4 days at the lower quintiles for Sukuk/Islamic equities in emerging markets.
However.
QC is positive across the median quintiles.
These results suggest that Sukuk is a haven asset and a diversifier for Islamic equities in emerging markets.
In addition.
Figure 5 shows that QC is positive across trading days at the lower quintiles for Sukuk/Islamic equities in the USA.
Also.
QC is positive across trading days at the median quintiles.
These results suggest that Sukuk is not a haven asset but a diversifier for Islamic equities in the US market.
Wavelet Quantile Correlation Gold/Equities (USA) 128-256 days Periods 64-128 days 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 2.
Wavelet Quintile Correlation (Gold/Equitie.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Wavelet Quantile Correlation Gold/Equities (Emerging Market.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 2.
Wavelet Quintile Correlation (Gold/Equitie.
(Continue.
Wavelet Quantile Correlation Sukuk/Equities (USA) 128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 3.
Wavelet Quintile Correlation (Sukuk/Equitie.
Diversifying Islamic Haven Assets Wavelet Quantile Correlation Sukuk/Equities (Emerging Market.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 3.
Wavelet Quintile Correlation (Sukuk/Equitie.
(Continue.
Markov-switching GARCH-Wavelet Quantile Correlation Gold/Equities (USA) 128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 4.
MSGARCH Ae WQC (Gold/Equitie.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Markov-switching GARCH-Wavelet Quantile Correlation Gold/Equities (Emerging Market.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 4.
MSGARCH Ae WQC (Gold/Equitie.
(Continue.
Markov-switching GARCH-Wavelet Quantile Correlation Sukuk/Equities (USA) 128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 5.
MSGARCH Ae WQC (Sukuk/Equitie.
Diversifying Islamic Haven Assets Markov-switching GARCH-Wavelet Quantile Correlation Sukuk/Equities (Emerging Market.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 5.
MSGARCH Ae WQC (Sukuk/Equitie.
(Continue.
TVGARCH Ae WQC Before estimating TVGARCH, nonstationary volatility must be detected.
Table 5 shows that the unconditional variance is not constant and has three transitions for the US equities.
The results indicate that the US equities are associated with TV.
Ae GARCH.
Moreover.
Table 6 presents the results from TV.
Ae GARCH.
of US equities.
Table 5.
Testing GARCH .
against TV Ae GARCH .
in the USA Panel A Ae US Equities Results from GARCH .
Estimate Std.
Error
Log-Likelihood
TR2
p-value Results from the Robust Test:
H0: c0 = c1 = c2 = c3 = 0 H03: c3* = 0 H02: c2* = 0 | c3* = 0 H01: c1* = 0 | c2* = c3* = 0 No.
of locations (=0.
= 3
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Table 5.
Testing GARCH .
against TV Ae GARCH .
in the USA (Continue.
Panel B Ae Gold Results from for GARCH .
Estimate Std.
Error Log-Likelihood Results from the Robust Test:
TR2
p-value TR2 p-value H0: c0* = c1* = c2* = c3* = 0 H03: c3* = 0 H02: c2* = 0 | c3* = 0 H01: c1* = 0 | c2* = c3* = 0 No.
of locations (=0.
= 0
Panel C Ae Sukuk Results from GARCH .
Estimate Std.
Error Log-Likelihood Results from the Robust Test:
H0: c0 = c1 = c2 = c3 = 0 H03: c3* = 0 H02: c2* = 0 | c3* = 0 H01: c1* = 0 | c2* = c3* = 0 No.
of locations (=0.
= 0
Notes: This table shows testing H0 .
0* = c1* = c2* = c3* = .
, which is the unconditional variance is time-invariant for equities (Panel A).
Gold (Panel B), and Sukuk (Panel C) in the US.
The results only reject Panel AAos null hypothesis .
-value is 0.
Table 6.
The estimation results of TV .
Ae GARCH .
Ae US Equities LongAeterm parameter .
ime-varying specificatio.
Size1 Speed1 Location1 Estimate Std.
Error Location2 Location3 ShortAeterm parameter (GARCH specificatio.
Estimate Std.
Error Log-Likelihood Notes: This table reveals the TV .
Ae GARCH .
parameters of US Equities.
Diversifying Islamic Haven Assets Similarly.
Table 7 indicates that the unconditional variance is not constant and has three transitions for the emerging market equities.
Thus, the results establish that the emerging market equities are also characterized by TV.
Ae GARCH.
Table 8 presents the results from TV.
Ae GARCH.
of Islamic equities in emerging markets.
Table 7.
Testing GARCH .
against TV Ae GARCH .
in Emerging Markets Panel A Ae Equities in Emerging Markets Results from GARCH .
Estimate Std.
Error Log-Likelihood Results from the Robust Test:
H0: c0 = c1 = c2 = c3 = 0 H03: c3* = 0 H02: c2* = 0 | c3* = 0 H01: c1* = 0 | c2* = c3* = 0 No.
of locations (=0.
= 3
TR2
Results from GARCH .
Estimate Std.
Error Log-Likelihood Results from the Robust Test:
H0: c0* = c1* = c2* = c3* = 0 H03: c3* = 0 H02: c2* = 0 | c3* = 0 H01: c1* = 0 | c2* = c3* = 0 No.
of locations (=0.
= 0
Results from GARCH .
Estimate Std.
Error Log-Likelihood Results from the Robust Test:
H0: c0* = c1* = c2* = c3* = 0 H03: c3* = 0 H02: c2* = 0 | c3* = 0 H01: c1* = 0 | c2* = c3* = 0 No.
of locations (=0.
= 0
p-value Panel B Ae Gold
TR2
p-value Panel C Ae Sukuk
TR2
p-value Notes: This table shows testing H0 .
0* = c1* = c2* = c3* = .
, which is the unconditional variance is time-invariant for US equities.
Gold, and Sukuk in the US.
The results only reject Panel AAos null hypothesis .
-value is 0.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Table 8.
The estimation results of TV .
Ae GARCH .
Ae Equities in Emerging Markets LongAeterm parameter .
ime-varying specificatio.
Size1 Speed1 Location1 Estimate Std.
Error Location2 Location3 ShortAeterm parameter (GARCH specificatio.
Estimate Std.
Error Log-Likelihood Notes: This table reveals the TV .
Ae GARCH .
parameters of US Equities.
Figure 6 reveals the conditional standard deviation for both a stationary GARCH.
and the TV.
Ae GARCH.
for US equities.
The figures show that the stationary GARCH .
understates the magnitude of the conditional standard deviation, especially during COVIDAe19.
Russia Ae Ukraine, and the Middle Eastern Similarly.
Figure 7 exhibits the conditional standard deviation for both a stationary GARCH.
and the TV.
Ae GARCH.
for emerging market The figures illustrate that the stationary GARCH.
also understates the magnitude of the conditional standard deviation in emerging market equities.
Similar to the results from MSGARCH-WQC.
Figures 8 and 9 reveal that gold and Sukuk are haven assets.
However.
MSGARCH-WQC indicates that Sukuk is not a haven asset for US equities, while TVGARCH-WQC reveals that Sukuk is a haven asset for US equities .
1 quintile of 128 Ae 256 days on Figure .
The results from TVGARCH are more robust since the TVGARCH package .
he R statistical progra.
has more features than MSGARCH, such as nonstationary testing, higher asymmetry order, and smooth transitions.
GARCH .
- USA
Figure 6.
Volatility based on GARCH.
and TV.
Ae GARCH.
of US Equities Diversifying Islamic Haven Assets TV .
- GARCH .
- USA Figure 6.
Volatility based on GARCH.
and TV.
Ae GARCH.
of US Equities (Continue.
GARCH .
- EMERGING MARKETS Figure 7.
Volatility based on GARCH.
and TV.
Ae GARCH.
of Equities in Emerging Markets Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 TV .
- GARCH .
- EMERGING MARKETS
Figure 7.
Volatility based on GARCH.
and TV.
Ae GARCH.
of Equities in Emerging Markets (Continue.
TV - GARCH-Wavelet Quantile Correlation Gold/Equities (USA) 128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 8.
TVGARCH Ae WQC (Gold/Equitie.
Diversifying Islamic Haven Assets TV - GARCH-Wavelet Quantile Correlation Gold/Equities (Emerging Market.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 8.
TVGARCH Ae WQC (Gold/Equitie.
(Continue.
TV - GARCH-Wavelet Quantile Correlation Sukuk/Equities (USA) 128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 9.
TVGARCH Ae WQC (Sukuk/Equitie.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 TV - GARCH-Wavelet Quantile Correlation Sukuk/Equities (Emerging Market.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 9.
TVGARCH Ae WQC (Sukuk/Equitie.
(Continue.
Simulation of TV Ae GARCH Sukuk ETF is a relatively new instrument in financial markets.
Hence, it is expected that the assets under the management of the Sukuk ETF will be greatly lower than those of the gold ETF.
This research simulates 10000 real equities, gold, and Sukuk returns to minimize size bias.
Figures 10 and 11 exhibit the simulated returns of the haven assets and Islamic equities in the US and emerging markets, respectively.
TV.
Ae GARCH.
parameters are used to simulate the inflation-adjusted returns of equities, while GARCH.
parameters are applied to simulate the inflation-adjusted returns of the haven assets.
Simply put, the parameters used for the simulation are from Tables 6 and 8 for Islamic equities in the US and emerging markets, respectively, and Tables 5 and 7 for the haven assets .
old and Suku.
in the US and emerging markets, respectively.
Based on the simulated returns, this research re-applies the new approach proposed in this study .
combination of TVARCH and WQC).
Figures 12 and 13 show the results.
Overall, the results reveal that gold and Sukuk are haven assets.
Portfolio Performance The dual momentum strategy has been explained in Section 3.
In the first part of this section.
We discuss the determination of the default weights of the strategy.
Figure 14 depicts the downside volatility and Sortino ratio throughout a wide range of Sukuk/Gold Ae Islamic equities allocations.
When combining 45% Gold or Sukuk .
hichever has the best performance in the last three month.
with 55% Islamic equities, the downside volatility is 0.
As expected, the downside volatility of this proportion is smaller than that of the Islamic equities .
ee Table .
due to the Diversifying Islamic Haven Assets safe-haven property of gold.
However, the reduction of the downside volatility comes at the cost of the Sortino ratio.
Thus, we use the following allocations for the dual momentum strategy: 55% Islamic equities Ae 45% Sukuk or Gold .
hichever has the best performance in the last three month.
uSim_ usa Simulated Real Retuns of US Equities Index uSim_ gold Simulated Real Retuns of Gold in USA Index Figure 10.
Simulated Real Returns in the USA Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 uSim_ Sukuk Simulated Real Retuns of Sukuk in USA Index Figure 10.
Simulated Real Returns in the USA (Continue.
uSim_ emerging Simulated Real Retuns of Equities in Emerging Markets Index Figure 11.
Simulated Real Returns in Emerging Markets Diversifying Islamic Haven Assets uSim_ gold Simulated Real Retuns of Gold in Emerging Markets Index uSim_ sukuk Simulated Real Retuns of Sukuk in Emerging Markets Index Figure 11.
Simulated Real Returns in Emerging Markets (Continue.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Simulation - TV - GARCH-Wavelet Quantile Correlation Gold/Equities (USA) 128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Simulation - TV - GARCH-Wavelet Quantile Correlation Gold/Equities (Emergin.
128-256 days 64-128 days Periods 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 12.
Simulation Ae TVGARCH Ae WQC (Gold/Equitie.
Diversifying Islamic Haven Assets Simulation - TV - GARCH-Wavelet Quantile Correlation Sukuk/Equities (USA) 128-256 days Periods 64-128 days 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Simulation - TV - GARCH-Wavelet Quantile Correlation Sukuk/Equities (USA) 128-256 days Periods 64-128 days 32-64 days 16-32 days 8-16 days 4-8 days 2-4 days Quantiles Figure 13.
Simulation Ae TVGARCH Ae WQC (Sukuk/Equitie.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Portfolio Performance (USA) Figure 15 depicts the portfolio weights.
The rebalancing period is quarterly instead of monthly to minimize turnover.
Yearly rebalancing is more economical but does not quickly adapt to market dynamics.
In addition, short-selling is not allowed since it is forbidden from an Islamic point of view.
The minimum variance approach allocates the majority of the assets to Sukuk.
Similarly, the passive and the risk parity methods allocate a large proportion of the portfolio to Sukuk.
Interestingly, the dual momentum approach indicates that investors should convert some portfolio assets to cash to avoid large drawdowns.
Further.
Table 9 shows that the dual momentum strategy outperforms the benchmark .
approach regarding risk-adjusted returns (Sharpe.
Sortino, and Omeg.
Additionally, the minimum variance strategy has the best Value-at-Risk at the cost of risk-adjusted returns.
Interestingly, the dual momentum strategy has the best drawdowns.
As expected, the all-equities portfolio has the worst Table 9.
Portfolio Performance .
he USA) Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % Naive Dual Momentum Min.
Variance Risk Parity All Equities Passive Notes: This table shows the portfolio performance.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more substantial the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
Figure 14.
The Determination of the Default Weights of the Dual Momentum Strategy Diversifying Islamic Haven Assets Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Asset allocation .
trategy: naive or the benchmar.
Date equities .
Asset allocation .
trategy: 30% Equities - 70% Sukuk or Passiv.
Date equities .
Figure 15.
Portfolio Weights .
he USA) Diversifying Islamic Haven Assets Asset allocation .
trategy: risk part.
Date equities .
Asset allocation .
trategy: minimum varianc.
Date equities .
Figure 15.
Portfolio Weights .
he USA) (Continue.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Asset allocation .
trategy: dual momentu.
Date equities .
Figure 15.
Portfolio Weights .
he USA) (Continue.
Portfolio Performance (Emerging Market.
Figure 16 depicts the portfolio weights for emerging markets.
The rebalancing period is quarterly.
Also, short-selling is not allowed.
The minimum variance approach allocates the majority of the time to Sukuk.
Compared to the dual momentum approach in the USA, the dual momentum strategy in emerging markets indicates that investors should retain more cash to avoid more significant It is a very reasonable action since the drawdowns in Islamic equities in emerging markets are more prominent .
ee Table .
Moreover.
Table 10 shows that the dual momentum strategy outperformes other strategies regarding risk-adjusted returns (Sharpe.
Sortino, and Omeg.
Additionally, the minimum variance strategy has the best Value-at-Risk.
Similarly, the 70% Sukuk Ae 30% Islamic equities portfolio has the second-best Value-at-Risk.
As predicted, the all-equities portfolio has the worst drawdowns.
At the same time, the dual momentum strategy has the best drawdowns.
Diversifying Islamic Haven Assets Table 10.
Portfolio Performance (Emerging Market.
Naive Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % Dual Min.
Momentum Variance Risk Parity All Equities Passive Notes: This table shows the portfolio performance.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more substantial the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are pretty confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
Asset allocation .
trategy: naive or the benchmar.
Date equities .
Figure 16.
Portfolio Weights (Emerging Market.
Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Asset allocation .
trategy: 30% Equities - 70% Sukuk or Passiv.
Date equities .
Asset allocation .
trategy: risk parit.
Date equities .
Figure 16.
Portfolio Weights (Emerging Market.
(Continue.
Diversifying Islamic Haven Assets Asset allocation .
trategy: minimum varianc.
Date equities .
Asset allocation .
trategy: dual momentu.
Date equities .
Figure 16.
Portfolio Weights (Emerging Market.
(Continue.
Robustness Tests The results in the previous sections are based on the followings: .
there is no transaction cost, and .
we use the entire data.
We re-evaluate portfolio performance pertaining to these two aspects Transaction Cost We impose a transaction cost of 63 basis points (Angel et al.
, 2.
and then reassess the portfolio performance.
Tables 11 and 12 show that the dual momentum strategy Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 still has the best drawdown, although the transaction cost slightly reduces the risk-adjusted returns.
Table 11.
Portfolio Performance .
he USA) Ae Net of Transaction Cost Nayve Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % Dual Min.
Momentum Variance Risk Parity All Equities Passive Notes: This table shows the portfolio performance.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more significant the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are pretty confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
Table 12.
Portfolio Performance (Emerging Market.
Ae Net of Transaction Cost Nayve Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % Dual Min.
Momentum Variance Risk Parity All Equities Passive Notes: This table shows the portfolio performance.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more significant the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are pretty confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
Diversifying Islamic Haven Assets New Data Partitioning Our earlier results are based on the entire sample, including low and high downturn periods.
Simply put, this exercise can help determine the long-term cost of administering a trading strategy.
Therefore, we would want a plan that does not suffer significant returns reduction during regular and bull markets .
y including haven assets in the equity portfoli.
while reducing downside risk during substantial market drawdowns.
For robustness, this study investigates the performance of the trading strategies during large drawdowns.
This research selects all periods in which equities experienced more than 10% drawdowns while maintaining a rolling window approach.
Figures 17 and 18 show the equities drawdowns throughout the years, highlighting the data partitioning.
The blue-shaded areas are used as a new sample.
Further.
Tables 13 and 14 indicate that the dual momentum strategy has the best drawdowns.
Figure 17.
New Data Partitioning (Emerging Market.
Figure 18.
New Data Partitioning .
he USA) Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Table 13.
Portfolio Performance .
he USA) Ae New Data Partitioning Naive Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % Dual Min.
Momentum Variance Risk Parity All Equities Passive Notes: This table shows the portfolio performance.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more significant the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are pretty confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
Table 14.
Portfolio Performance (Emerging Market.
Ae New Data Partitioning Naive Mean returns % Std.
Downside Volatility % Minimum returns % Maximum returns % Skewness Kurtosis Sharpe % Sortino % Omega Value-at-Risk % Median drawdown % Maximum drawdown % Dual Min.
Momentum Variance Risk Parity All Equities Passive Notes: This table shows the portfolio performance.
A positive skewness .
typically has minor losses and few significant gains.
The larger the risk-adjusted returns (Sharpe.
Sortino, and Omeg.
, the better the performance.
The larger the drawdown, the more substantial the decrease in an investmentAos peak-to-trough over time and the riskier the investment.
Similarly, the larger the downside volatility, measured by semi-deviation, which assesses the below mean fluctuations, the riskier the investment.
In addition, the lower the value-at-risk, a loss that we are pretty confident will not be surpassed if the portfolio is held for a certain period, the riskier the investment.
Diversifying Islamic Haven Assets Analysis/Discussion This section discusses the new approach and the alternative version of the dual momentum strategy.
It then discusses the Sukuk performance for hedging compared to previous studies, which is still lacking in the current literature.
In the last part of this section, we discuss the importance of this research to the literature on Islamic finance.
Table 1 presents the limitations of prior studies.
This research relaxes the assumptions made in the previous studies.
GARCH-based methods and quintile regression are widely used in the literature to evaluate the haven qualities of gold and Sukuk.
However, those methods have limitations.
For example, the quintile regression based on the dynamic correlation of DCCAiGARCH does not show the haven qualities of haven assets at different quintiles.
Thus.
Kumar & Padakandla .
have developed the Wavelet Quintile Correlation (WQC) method to address the limitations of the methods used in the previous studies.
This study extends the Wavelet Quintile Correlation (WQC) method of Kumar & Padakandla .
, incorporating nonstationary volatility.
Specifically, this research uses the current state-of-the-art methodologies, including MSGARCH and TVGARCH.
A dual-regime volatility is statistically more appropriate than a single-regime one.
Furthermore, the TVGARCH model is also employed to enhance the quality of WQC.
Additionally, this study enhances the original WQC by using the inflation-adjusted returns, avoiding the illusion of money.
This study shows that Sukuk can be a haven asset.
This finding contrasts with the previous studies that show Sukuk is not a haven asset (Naeem et al.
, 2023.
Nugroho & Kusumawardhani, 2023.
Qadri et al.
, 2.
The possible reason is that previous studies do not consider volatility dynamics.
Figure 19.
Drawdowns of Dual Momentum Versus Constant Weight Journal of Islamic Monetary Economics and Finance.
Vol.
Number 3, 2025 Further, this study develops another version of the dual momentum strategy based on the results from MSGARCH-WQC.
We rebalance the portfolio quarterly to adapt to market dynamics while having a moderate portfolio turnover.
This strategy provides a better drawdown than the constant weight strategy with yearly rebalancing .
% Gold, 45% Equities, 10% Gol.
proposed by Vliet & Lohre .
The transaction cost is 63 basis points.
Moreover, this research extends Masih et al.
review of the current quantitative studies on Islamic equities, such as the performance of Islamic equities relative to non-Islamic ones and Socially Responsible Investing or SRI (Abdelsalam et al.
, 2014.
Charfeddine et al.
, 2016.
Charles et al.
, 2015.
Jawadi et , 2.
, the performance of Islamic equities during a bear market (Ajmi et al.
, the diversification benefits of Islamic equities in international portfolio (Majdoub & Mansour, 2.
, the analysis of Islamic equities risk (Bekri & Kim, 2.
, calendar anomalies (Abbes & Abdelhydi-Zouch, 2.
, and the overview of Shariah-compliant equities screening parameters (Clarke, 2.
CONCLUSION AND RECOMMENDATION
Conclusion This research modifies the Wavelet Quintile Correlation (WQC) method and then develops a trading strategy involving Islamic assets (Gold.
Sukuk, and Islamic Mainly, this study combines nonstationary volatility models (Markovswitching GARCH and Time-varying GARCH) with the conventional WQC to allow for the volatility dynamics.
The daily data are four exchange-traded funds:
Dow Jones Global Sukuk.
Wahed FTSE USA Shariah.
MSCI Emerging Market Islamic, and SPDR Gold.
The return series are adjusted to inflation to avoid the money illusion.
The results show that Sukuk and Gold are haven assets.
Next, this study optimizes the SukukAiGoldAiIslamic equities portfolio.
implement several widely used methods, such as nayve .
he benchmar.
, minimum variance, risk parity, passive investing .
% SukukAi30% Islamic equitie.
, and the trend-following strategy .
ual momentu.
The rebalancing frequency is quarterly to minimize transaction costs.
The results show that the dual momentum strategy has the best drawdown for extreme declines in the equities markets.
Recommendation The results of this research can be beneficial for both academics and investors.
This research enhances the WQC model recently proposed by Kumar & Padakandla .
For investors, the recommended trading strategy for the Sukuk Ae Gold Ae Islamic equities portfolio is the dual momentum model with quarterly rebalancing.
However, transaction costs may greatly impact portfolio performance.
The proposed strategy is only preferable when the transaction cost is below 70 basis points .
he average transaction cost for ETF is 63 basis point.
The naive strategy is suggested when the transaction cost exceeds 70 basis points.
In other words, this research suggests assigning a relative value to each asset class (Sukuk.
Gold, and Islamic equitie.
based on how well it has performed over the past three months relative to other assets in the same class (Sukuk and Gold are havens, in Diversifying Islamic Haven Assets the strategy they are in the same clas.
and whether or not it has had a positive As long as the top-performing asset in the asset class has a positive return above zero, dual momentum invests in those assets.
Otherwise, the allocation is shifted to cash.
REFERENCES