Unpacking the financial attributes of blue-chip non-fungible tokens (NFTs) against traditional and digital assets

Abstract

Purpose

This study aims to identify the financial attributes of non-fungible tokens (NFTs) as safe havens, hedges or diversifiers against traditional (stock indices, foreign exchange, gold and government bonds) and digital (Bitcoin and Ethereum) assets.

Design/methodology/approach

The quantile via moments was utilized, and the data spanned from 20 September 2021 to 31 January 2022. The authors incorporated feasible generalized least squares (FGLS) and difference-generalized method of moments (diff-GMM) as the robustness check.

Findings

Overall, NFTs offer strongly safe havens, hedging and diversifier attributes against cryptocurrencies, while weak properties for traditional assets. The specific findings are: (1) Bored Ape Yacht Club (BAYC) serves as a strong hedge for Bitcoin during market rise; (2) Mutant Ape Yacht Club (MAYC) serves as a strong safe haven against Bitcoin during market bull; (3) Crypto punk (CP) provides strong safe havens properties for gold during market turmoil while serving as a strong hedge against gold and Bitcoin on average and (4) the three blue-chip NFTs are powered by Ethereum blockchain, thus serving as a diversifier against Ethereum.

Practical implications

Bitcoin investors are suggested to include NFTs in their investment portfolio to mitigate the losses when Bitcoin falls. Meanwhile, the inclusion of crypto punk is advised for risk-averse investors who invest in gold. NFTs are powered by the Ethereum blockchain, indicating co-movement among them and thus, serve as diversifiers. Policymakers and regulators are suggested to watch closely over NFTs' great development and restructure the existing policies and thus, stabilization of asset markets can be achieved.

Originality/value

The originality aspects are: (1) focusing on the three blue-chip NFTs (i.e. BAYC, MAYC and CP) that are categorized as the largest NFTs by floor market capitalization; (2) testing the NFT attributes (safe havens, hedges or diversifiers) against traditional and digital assets, a.k.a., cryptocurrencies and (3) panel setting on 14 countries with the highest NFT users.

Citation

Havidz, S.A.H., Santoso, M.D., Alexander, T. and Caroline, C. (2024), "Unpacking the financial attributes of blue-chip non-fungible tokens (NFTs) against traditional and digital assets", Asian Journal of Accounting Research, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/AJAR-10-2023-0334

1. Introduction

Although digital assets such as Bitcoin and Ethereum show rapid development (Zhang et al., 2022) due to the blockchain technology first established by Nakamoto (2008), non-fungible tokens (NFTs) provide blockchain technology novelty (Wilson et al., 2022). Public interest in blockchain technology is growing rapidly. In 2016, over $1bn was invested in blockchain solution deployment held by financial services and technology companies to their businesses, and it is expected to increase dramatically in the next few years [1]. Blockchain technology, which delivers trust-free transactions without intermediaries, has been an interest since 2008 during the financial crisis that happened globally. It occurred due to the financial institutions and banks who failed to mitigate risk. Blockchain is believed to be a solution to the trust deterioration in intermediaries, especially financial intermediaries, since it eliminates the need for third parties (intermediaries) when doing transactions (De Filippi et al., 2020). It can be seen from global spending on blockchain solutions rising from $0.95bn in 2017 to over $4.5bn in 2020. Statistics also forecast that over the year 2021 until 2024, the spending will continue to grow from 6.6bn US dollars to reaching almost $19bn in 2024 [2].

At the beginning of March 2021, the new existence of NFT art entitled Everyday: The First 5000 Days sold 69m US dollars and became the largest art sale in history (Rehman et al., 2021). Since then, public attention on NFT exploded and made OpenSea one of the largest NFT marketplaces, reaching one million users in early 2022 from only 37,400 users in 2021 [3]. NFTs are cryptographic assets that produce images, videos, songs and virtual lands (Kanellopoulos et al., 2021) and offer potential revenue and store value (Kumar and Padakandla, 2023). Not only obtaining social attention (Zhang et al., 2022) but also young investors are very excited about the revolution of new digital assets, a.k.a., NFTs. Shortly, it could be a prospective digital asset (Ali et al., 2023a).

Every day, thousands of hundreds of NFTs were traded with an average volume of $3bn per day [4] and it was proven that NFTs serve not as currency, a commodity or a technology, but as an asset (Dowling, 2022). Although NFTs are difficult to value, it is observed that NFTs are a new asset class that may mature over time. Shortly, NFTs can be strongly related to cryptocurrency as its native currency and other asset classes (Ante, 2021). These indicate that NFTs can be used as an alternative investment.

Financial researchers have examined the connectedness and spillover studies of NFTs with existing financial assets or among NFTs (Dowling, 2022; Aharon and Demir, 2022; Urom et al., 2022; Umar et al., 2022a, b). Remarkably, we found scarce research that focuses on the safe havens and hedge attributes of NFTs (Zhang et al., 2022; Ko and Lee, 2023; Kumar and Padakandla, 2023). Therefore, we are motivated to identify the function of NFTs, either as safe havens, hedges or diversifiers against financial assets. We examine a comprehensive study by utilizing the blue-chip NFTs against traditional (i.e. stock, foreign exchange, gold and government bond) and digital (i.e. Bitcoin and Ethereum) assets. Such work remains unexplored.

Zhang et al. (2022) examined the NFTs' attributes as safe havens and hedges against the US dollar, bonds, stocks, crude oil, gold and Bitcoin. Ko and Lee (2023) utilized the NFT index to discover either as a safe haven or hedge against the financial asset indices (i.e. stock, commodity, bond and US dollar) and cryptocurrencies (i.e. Bitcoin and Ethereum). Kumar and Padakandla (2023) employed log returns from four NFTs average prices (i.e. Cryptokiteis, Decentraland, SuperRare and Cryptopunks) and further examined the safe havens or hedge properties against Bitcoin and Ethereum. None of the earlier studies focused on the NFTs individually, while investors are going to invest in certain NFTs that they have an interest in. Therefore, we fill the research gap by examining the NFT attributes individually on the three blue-chip NFTs (i.e. Bored Ape Yacht Club (BAYC), Mutant Ape Yacht Club (MAYC) and Crypto-Punks (CP)).

Our research contributes by providing new insights to the literature, investors and stakeholders in threefold (1) updating the function and effectiveness (i.e. safe havens, hedges and diversifiers): literature, as the prior works favor gold as safe haven assets (Shahzad et al., 2020; Li and Lucey, 2017; Baur and Lucey, 2010; Lucey and Li, 2015; Baur and Mcdermott, 2010). Cryptocurrencies (Bitcoin and Ethereum) as digital assets have also been examined (Antonakakis et al., 2019; Disli et al., 2021; Stensås et al., 2019; Mariana et al., 2021; Kang et al., 2020). Through this study, the revolution of NFTs as new digital assets can be determined; (2) we selected the sample based on the largest users of NFT based on Statista (2022) and thus, panel setting is applied (3) focusing on the three blue-chip NFTs (i.e. BAYC, MAYC and CP) that are categorized as the largest NFTs by floor market capitalization.

The rest of the section on this paper is structured as follows. Discusses tokens, financial attributes theory and related prior works in section 2. Section 3 explains the data and core methodology. Section 4 elaborates on the findings. Discusses further analysis of the results in section 5. Lastly, conclusions and recommendations in section 6.

2. Literature review

2.1 Tokens

The application of blockchain-based has gained popularity among Fintech breakthroughs (Cong et al., 2021). Blockchain technology provides smart solutions for traditional platforms that heavily depend on the innovations of payment to stimulate the exchanges of economics among users. The birth of digital platforms offers tokens of crypto to serve as local currency (Cong et al., 2022). Not only as currency but also tokens are considered as investment tools. Asset-backed tokens rely on an asset to back and thus have intrinsic value (Hassan et al., 2021). Gold-backed cryptocurrency (GBC) garnered the attention of investors since the beginning of COVID-19 (Jalan et al., 2021). Hassan et al. (2021) explored the precious metal-backed cryptocurrency (PMBC), while gold-backed Islamic cryptocurrencies (GBIC) provide new attractive investment tools to faith-based investors (Lahmiri and Bekiros, 2019; Aloui et al., 2021; Mnif and Jarboui, 2021, 2022; Ali et al., 2022, 2024; Nugroho, 2023; Trichilli and Boujelbéne, 2023; Zaman et al., 2023). The energy sectors have also adopted the digital revolution by transforming to energy tokens, which are considered unconventional instruments of financial. This has garnered the attention of commodity investors (Ali et al., 2023b; Yousaf et al., 2022). Tokenization indeed provides better services in the form of non-fungible tokens (NFTs) and Decentralized Finance (DeFi) (Umar et al., 2022a, b; Yousaf and Yarovaya, 2022; Corbet et al., 2023), asset management (Yousaf et al., 2023) and metaverse (Vidal-Tomás, 2022) tokens.

2.2 The theory of safe havens, hedges and diversifiers

The Modern Portfolio Theory is an investment method for investors to have an efficient portfolio by diversifying the assets with expected return and variance of return consideration; thus, the return will be optimized and the risk will be tolerable (Markowitz, 1952). To make an efficient portfolio, correlation among securities is one of the most important aspects, which needs to be non-correlated and low-correlated with other securities to eliminate and reduce the risk (Markowitz, 1971). Baur and Mcdermott (2010) and Baur and Lucey (2010) define an empirical testable of the three types of asset diversification in an investment portfolio (i.e. safe havens, hedges and diversifiers). A safe haven is a property that is negatively correlated to other financial assets in market turmoil. A hedge asset is a property that is negatively correlated to other financial assets on average. Meanwhile, a diversifier is positively correlated to other financial assets on average.

2.3 Previous research on safe havens, hedges and diversifiers

The literature on NFTs as safe havens and hedges has been growing but is still limited. Kumar and Padakandla (2023) confirm that NFTs are exhibited as a hedge against Bitcoin by employing the effectiveness of a dynamic hedge. Further, based on the Wavelet Quantile Correlation results, NFTs offer a safe haven to Bitcoin for the short run. Meanwhile, NFTs serve as diversifiers for Ethereum in the short-medium term. Ko and Lee (2023) show that NFTs are attributed as safe havens and hedges against certain stocks, oil, US indices and bonds with various degrees. It is further confirmed during the outbreak of COVID-19 that NFTs provide safe haven attributes for USD indices and bonds. Zhang et al. (2022) employ the NARDL model and find a hedge property of NFTs against bonds, gold and USD on normal market for the full period. Before COVID-19, NFTs had hedge benefits for the US dollar and stocks on average, while safe havens against USD during the outbreak.

Before NFTs gained popularity, numerous studies tested how gold, cryptocurrencies (Bitcoin and Ethereum) and other financial assets functioned against financial assets. Baur and Lucey (2010) find gold provides a safe haven attribute to the UK, US and German stock indices. Not only serving as a safe haven for conventional assets but also gold is able to be a safe haven property for Islamic equities (Tirosh and Barkai, 2007; Chkili, 2017). Lucey and Li (2015) applied multivariate GARCH DCC and proved that gold acted as a safe haven against US bonds and stocks. Li and Lucey (2017) further reassess and find gold acts as a safe haven for bonds (France, Canada, Italy and the UK) and stocks (Germany, the US, Italy, Japan and the UK). Shahzad et al. (2019) depict weak gold as safe haven property in emerging and developed stock markets. Indeed, the superiority of gold as a safe haven was confirmed during the COVID-19 crisis and uncertainty (Ji et al., 2020; Liu, 2020; Huang and Chang, 2021). Kumar and Padakandla (2022) indicate that gold remains to act as a safe haven (hedges) in the long and short (medium) term. Widjaja et al. (2023) evidence that gold comes out as the big winner of safe havens against stocks and bonds in Islamic and conventional markets. Meanwhile, the UCRY price only serves as a strong safe haven for conventional equities, while weak safe haven attributes for Islamic stocks.

Before the COVID-19 period, Bitcoin was a poor hedge except for weekly movement in China stocks. US dollars, gold, commodities and Bitcoin can be safe haven assets against stock markets (Wen and Cheng, 2018; Chen and Wang, 2019). Nevertheless, it can give a diversification benefit against almost all the stock indices, US dollar, commodities, oil and gold, while safe-haven benefits against Asian stocks (Bouri et al., 2017). The hedge and diversifier attribute of Bitcoin against several financial assets is confirmed, while a safe haven benefits against the monetary market in China (Wang et al., 2019). Bitcoin acts as a weak hedge property to Estonia, Sweden, Venezuela, China and Japan stocks in US dollars trade. Meanwhile, in local currency trade, it serves as a safe haven property in Venezuela and also a diversifier property in China and Japan (Kliber et al., 2019). Bitcoin provides a safe haven benefit to US stocks and a diversifier in the long run. Bitcoin’s utility to be a diversifier is confirmed, despite it not serving as a safe haven for other financial assets (Lavelle et al., 2022).

During the COVID-19 pandemic, it is further examined and proven Bitcoin serves as a safe haven for all sectors of US stocks, while Ethereum serves as a hedge only toward several sectors (Bouri et al., 2020a). Moreover, Bitcoin also has a hedge potential against stock prices but is a diversifier during extreme market conditions (Garcia-Jorcano and Benito, 2020). Several cryptocurrencies, including Bitcoin, are able to give diversification gains within a short-term period (Corbet et al., 2018) and safe haven benefits (Bouri et al., 2020b). Bitcoin also remains a diversifier (Huang et al., 2021). By contrast, Bitcoin and Ethereum cannot serve as a safe haven for stock markets due to increasing the downside risk portfolio (Conlon and McGee, 2020; Conlon et al., 2020).

3. Data and methodology

3.1 Data

We selected three blue-chip NFT projects with the biggest capitalization of market as of February 2022 (i.e. Bored Ape Yacht Club (BAYC), Mutant Ape Yacht Club (MAYC) and Crypto-Punks (CP)) that are operated by using the Ethereum chain [5]. BAYC is an ape-type NFT project that is being developed by Yuga Labs. The lucky holders will receive a reward for having a mutation version of their Bored Ape, known as Mutant Ape. MAYC is created by serums that had been airdropped randomly to the Bored Ape token holders. It is a form of the final tier of BAYC membership [6]. Meanwhile, Crypto-Punks is a character type of NFT project that was created by Larva Labs and has been acquired by Yuga Labs [7].

To provide reliable results, we focused on the countries with the largest users of NFT based on Statista (2022) with the criteria as listed in Table 1. As 14 countries were selected as our samples, we covered the daily period from 20 September 2021 to 31 January 2022. The data with five days of daily data basis (i.e. stock indices, foreign exchange and government bonds – see Table A1 in the appendix) were linearly interpolated to seven-day daily data. Linear interpolation is the simplest way of interpolation method that allows the estimation of missing values when there is a gap between the measurements (Gray et al., 2018). Hence, the entire variables are using seven days of daily data. We used nine indicators, which consist of three dependent variables (bored ape, mutant ape and crypto punk) and six independent variables (stock indices (SR), foreign exchange (FR), gold (GR), government bonds (GBR), Bitcoin (BTC) and Ethereum (ETH)). Three variables (stocks, foreign exchange and government bonds) employed country-level data, while the remaining variables employed global data. As for the independent variables, we extract the return data from the sources, while daily closing price data for the dependent variables and then, transform it into a return form, which is calculated as (P₁−P₀)/P₀ ✕ 100. Therefore, the entire variables are in the return form. Data sources and summary variables are attached in Table 2.

3.2 Quantile regression via moments

Koenker and Bassett (1978) first introduced regression quantiles further developed by Gutenbrunner and Jureckova (1992), which consider the conditional means. Following Dias et al. (2022), Gozgor and Karakas (2023) and Raifu et al. (2023), we also employ the quantile regression constructed by Machado and Santos Silva (2019), named quantile via moments to identify the safe haven, hedge, and diversifier attributes of NFTs against traditional and digital assets. Some benefits of applying quantile via moments are: (1) a novel approach of quantile regression to accommodate the heterogeneous issue of panel data, in our case the countries (Raifu et al., 2023); (2) it addresses a new way of predictors that impact the whole conditional distribution, especially when the predictors deal with the endogenous issue and entrench with individual effects (Awosusi et al., 2022). The equation of our quantile regression is presented as follows:

(1)Qy (τ|Xit)=(ai(τ)+δiq(τ))+Xit′β(τ)+Zit′ γ(τ)

Where Q_y(τ|X_it) describes the quantile for dependent variables, (a_i(τ) reflects the quantile of the distributional effect, which refers to countries used in this paper, δ_iq(τ) denotes the scale effect, X′_itβ(τ) implies the independent variables and Z′ refers to the known differentiable vector transformations of X components. Given the equation, as provided above, our model equation can be formed as follows:

(2)(BAYC|MAYC|CP) Qy (τ|Xit)=(ai(τ)+δiq(τ))+β1SRit +β2FRit +β3GRit +β4GBRit +β5BTCit+β6ETHit +Zit′γ(τ)

Referring to Yıldırım and Çelik (2021) and Singh and Kannadhasan (2020), we applied 10%–90% quantiles. We classify the nine quantiles into three phases based on market conditions that reflect the NFTs market condition: (a) bearish market (10–30%), (b) normal market (40–60%) and (c) bullish market (70–90%). The classification is divided into three: (1) strong (weak) safe haven is classified when revealing a negative coefficient and statistically significant (insignificant) in a bearish market, (2) strong (weak) hedge is classified when revealing a negative coefficient and statistically significant (insignificant) in normal and bullish markets and (3) strong (weak) diversifier is classified when revealing a positive coefficient and statistically significant (insignificant) under all market conditions.

To ensure the quality of the data, we ran several tests before proceeding to the quantile regression test. The cross-section dependency test by Pesaran (2021) was conducted first to check the dependency between the cross-section data used in this research. Second, we run a panel unit root test by Levin et al. (2002) to ensure the stationarity in our time series data. We used the LLC method because all variables are using the same number of time series (Choi, 2001).

3.3 Robustness tests

We conducted two-fold robustness tests as additional regressions to check the consistency results from our main regression. Feasible generalized least square (FGLS) was employed in the first step. Due to the existence of heteroskedasticity, the FGLS model is more efficient than OLS (Bai et al., 2020). Not only solving the heteroskedasticity issues but also FGLS is more attractive for large observations (Wooldridge, 2018). Arellano and Bond (1991) found difference generalized methods of moments (diff-GMM), which we employed on the second stage of robustness. The benefits of applying diff-GMM are overcoming the endogeneity from the regressors, and it can avoid dynamic panel bias (Abdeljawad et al., 2013).

4. Results

4.1 Unit root test and cross-section dependency results

The findings show that all variables utilized are stationary at level, which means there was no unit root problem in our time series data. The cross-section dependency test shows the existence of cross-section dependency between countries used in this paper. Therefore, the null hypothesis, which refers to no cross-sectional dependence, was rejected at 1% significance for all variables. Refer to Table 3 for both results.

4.2 Quantile via moments results

As for comparison, ordinary least square (OLS) is included, and the results implied correspond directions with quantile regression. Refer to Table 4. Panel A: BAYC, we found positive [negative] BAYC coefficient and statistically insignificant to stock under bearish and bullish (10–30% and 80–90%) [normal] (50–60%) conditions. It reveals that BAYC served as a weak diversifier [weak hedge]. The negative coefficient of BAYC statistically insignificant to foreign exchange, gold and government bonds under all market conditions signifies that BAYC offered a weak safe haven and hedge. Our findings support the prior findings by Zhang et al. (2022), Aharon and Demir (2022), Umar et al. (2022). Confirming the findings of Kumar and Padakandla (2023), a negative BAYC coefficient and statistically insignificant [significant] against Bitcoin under bearish and normal (10–50%); [bullish] (70–80%) conditions. It implies BAYC serves as a weak safe haven and hedge [strong hedge] against Bitcoin. BAYC offered a strong diversifier for Ethereum in all market conditions (10–40% and 60–80%) as it shows a positive coefficient of BAYC and is statistically significant as argued by Kong and Lin (2021) and Dowling (2022).

Panel B: MAYC of Table 4 reveals a positive coefficient of MAYC and is statistically insignificant to stocks under all market conditions (10–20%; 40–50% and 70–90%). It implies that MAYC serves as a weak diversifier for stocks. As for foreign exchange and gold, it was found that the MAYC coefficient outlines a positive significant [insignificant] in bear [normal] markets (10–30%) [40–60%]. It implied that MAYC acted as a strong [weak] diversifier for foreign exchange and gold under both markets mentioned earlier. Our results support the earlier findings by Kong and Lin (2021), which confirm the function of NFTs as diversifiers. Meanwhile, MAYC proved as a weak diversifier [strong hedge] in the bull market against foreign exchange [gold], as the coefficient of MAYC reveals positive insignificant [negative significant]. MAYC exhibited a negative coefficient and was insignificant against government bonds in all markets (10–90%). This signifies MAYC acted as a weak safe haven and hedge against bonds as argued by Ko and Lee (2023) and Zhang et al. (2022). Supporting the findings of Kumar and Padakandla (2023), MAYC revealed a negative coefficient and significance during market crises (20–30%) for Bitcoin, which serves as a strong safe haven. Meanwhile, a positive, insignificant coefficient of MAYC against Bitcoin on average (40–50% and 80–90%) refers to a weak diversifier. A negative insignificant coefficient of MAYC against Ethereum on average (40–60%) proved a weak hedge, while a strong diversifier of MAYC against Ethereum during the market rise and sound economy (80–90%). We contribute new findings to the earlier findings of Kumar and Padakandla (2023) that NFTs could serve as a safe haven and hedge not only for Bitcoin but also for Ethereum. Additionally, diversifier properties of NFTs not only served for Ethereum but also for Bitcoin. In this case, MAYC is the NFT.

Table 4. Panel C: crypto punk exhibited a negative insignificant [positive insignificant] coefficient of CP against stocks during market crises (10–20%) [normal market and rise economy] (40–50% and 70–90%). Hence crypto punk could act as a weak safe haven and diversifier for stocks (Kong and Lin, 2021; Ko and Lee, 2023). Moreover, crypto punk also serves as a weak diversifier at all market conditions against foreign exchange and government bonds because it shows positive, insignificant coefficients of crypto punk. A negative significant coefficient of crypto punk against gold at the bear market (10–30%) proves strong safe haven attributes of crypto punk against gold, while crypto punk serves as a strong hedge on average for gold and Bitcoin because of the negative significant coefficient of crypto punk at the normal market. Similar to BAYC, crypto punk offers strong diversification benefits during the market downturn for Ethereum because of the positive coefficient and significance of crypto punk against Ethereum. On the other hand, it reveals a weak diversifier on average by the positive, insignificant coefficient of crypto punk against Ethereum. Therefore, our new findings indicate that crypto punk offered diversification benefits not only for Ethereum (Kong and Lin, 2021; Dowling, 2022) but also for foreign exchange and government bonds.

4.3 FGLS and diff-GMM as robustness check

The robustness test results from FGLS and diff-GMM revealed consistency directions. Therefore, our estimations were robust across two different approaches among the regressors toward the NFTs variables (refer to Table 5).

5. Discussion

This research utilized three blue-chip NFTs (i.e. bored ape, mutant ape and crypto punk) (Bloomberg News, 2023a, b) and further examined the financial properties. NFTs, as a blockchain-based technology, could only serve as a strong hedge and safe haven for digital assets that also applied the blockchain system, Bitcoin. Bored ape and crypto punk revealed a strong hedge attribute for Bitcoin on average, while mutant ape offers strong safe haven benefits to Bitcoin during market turmoil. Our results supported the findings of Kumar and Padakandla (2023). Bitcoin investors could mitigate the risk by including mutant ape during the market downturn, while bored ape and crypto punk during the economic recovery or normal market to the investment portfolio. Similarly, the NFTs market is dominated by risk-seeking investors who have a high desire to build a high-risk portfolio (Hayes, 2022). The large, sudden numbers of investors entering the NFT markets have led to a bubble coin. Given that situation, it creates higher volatility among cryptocurrencies, thus raising their values (Wilson et al., 2022).

Confirming the results of Zhang et al. (2022), NFTs, in our case crypto punk, also serve as a strong hedge against gold. However, we provide new insights into the safe haven benefits of crypto punk against gold. Therefore, crypto punk is the only NFT that could serve as a safe haven and hedge to commodity assets. In addition, crypto punk, as one of the pioneers in NFT [8], has the holders' behavior tending to hold rather than pump and dump, which could maintain the market stability that will impact its risk. Therefore, it may attract investors who are risk-averse and are inclined to invest in gold to make a less risky portfolio (Wong, 2009).

Although Bored Ape also offers safe haven and hedge benefits against traditional assets (i.e. stocks, foreign exchange, gold and government bonds), it is only in a weak mode. Hence, it is insignificant to include bored ape in the portfolio when having traditional assets. In contrast, mutant ape offers a strong diversification for foreign exchange and gold during market downturns and, on average, is weak for stocks. Confirming the findings of Ko and Lee (2023), NFTs, in our case mutant ape, could absorb the unexpected shocks and thus provide diversification opportunities against traditional assets. Further, mutant ape also provide safe havens and hedge property against government bonds but are weak. Overall, the three blue-chip NFTs are powered by the Ethereum blockchain in conducting the transactions, then it serves as a diversifier against Ethereum under the majority of market conditions. The results are consistent with Kumar and Padakandla (2023), exhibiting a co-movement between Ethereum and NFTs.

6. Conclusion and recommendations

NFTs can be seen as potential alternative investments. Given the situation that there are more upcoming people, begin to see and accept NFTs as a new digital asset. These findings support Ante (2021) that NFTs will mature over time. Our paper by far gives the most relevant results by using data from selected 14 countries with the highest number of NFT users.

We provide economic implications to policymakers, academic researchers, regulators and investors. NFTs are indeed attractive during the new entrance, which is also largely endorsed by celebrities, thus sharply raising the value. Yet, the NFTs environment is highly volatile and infamously unpredictable (Bloomberg News, 2023c). Therefore, the government and regulator play an important role in determining the future of NFTs. The policy of the government and regulator will drive investors’ decision-making and behavior in choosing digital assets. We suggest that governments and regulators should thoroughly analyze the potential of NFTs, construct proper regulations and monitor NFT development because NFTs are potentially having great development (Zhang et al., 2022). Notably, NFTs can be a shelter for investors to diversify their portfolios and manage risk. Our research can be further explored that account for unprecedented times.

Table 1

Selection criteria

No	Criteria	Countries
1	Countries with the highest number of adults who either own or plan to buy NFT	19
2	Countries with unavailable forex data	1
3	Countries with unavailable 10 years of government bonds data	4
	Countries selected as the samples	14

Source(s): Created by authors

Table 2

Descriptive summary and data sources

Variables	Sources	Mean	Std dev	Min	Max	Obs
BAYC	Coingecko.com	0.85	4.58	−10.96	19.82	1,876
CP	Coingecko.com	−0.59	3.92	−8.22	21.06	1,876
MAYC	Coingecko.com	1.67	9.33	−8.52	86.25	1,876
SR	Investing.com	0.23	1.04	−4.15	6.31	1,876
FR	Investing.com	−0.14	0.43	−3.05	2.37	1,876
GR	Investing.com	−0.17	0.63	−2.43	2.02	1,876
GBR	Investing.com	0.20	9.47	−129.79	200	1,876
BTC	Investing.com	−0.09	3.33	−10.41	9.86	1,876
ETH	Investing.com	−0.07	4.19	−14.43	11.45	1,876

Source(s): Created by authors

Table 3

Results from LLC-test and CD-test

Variables	LLC-test	CD-test
BAYC	−23.2212***	110.426***
CP	−22.5191***	110.426***
MAYC	−24.1667***	110.426***
SR	−18.4437***	26.693***
FR	−19.2159***	13.277***
GR	−20.5194***	110.426***
GBR	−16.4202***	11.476***
BTC	−30.6115***	110.426***
ETH	−32.5763***	110.426***

Note(s): ***Stands for 1% level of significance

Source(s): Created by authors

Table 4

Ordinary least square and quantile regression results

	SR	FR	GR	GBR	BTC	ETH
Panel A: BAYC
OLS	0.127	−0.199	0.276*	−0.011	−0.078	0.153***
0.1	0.002	−0.02	0.508*	−0.003	−0.158	0.241***
0.2	0.005	−0.204	−0.202	−0.011	−0.099	0.267***
0.3	0.063	−0.242	−0.1	−0.011	−0.023	0.155***
0.4	0.013	−0.313**	−0.097	−0.003	−0.042	0.115***
0.5	−0.022	−0.128	−0.066	−0.002	−0.077	0.085
0.6	−0.03	−0.12	−0.256	−0.013	−0.353***	0.279***
0.7	−0.042	−0.269	0.479**	−0.024	−0.255***	0.242***
0.8	0.09	0.033	−0.107	−0.013	−0.422***	0.319***
0.9	0.489	−0.592	−0.232	−0.017	−0.031	−0.066
Panel B: MAYC
OLS	0.08	1.298***	−0.143	−0.032	−0.132	−0.077
0.1	0.172	1.401***	1.082***	−0.015	−0.162	−0.075
0.2	0.216	0.793**	0.841***	−0.01	−0.326***	0.151**
0.3	0.441**	0.84*	0.534*	−0.013	−0.268***	0.102
0.4	0.127	0.262	0.165	−0.01	0.054	−0.034
0.5	0.008	0.042	0.353***	−0.005	0.042	−0.008
0.6	−0.063	0.233	0.319	−0.023	0.249***	−0.054
0.7	0.116	0.584	−0.234	−0.029	0.394***	−0.109
0.8	0.146	0.073	−1.167***	−0.02	0.041	0.178**
0.9	0.243	0.366	−2.875***	−0.026	0.034	0.449**
Panel C: CP
OLS	0.017	0.547***	−0.355**	0.008	0.048	0.054
0.1	−0.113	0.369	−0.661***	0.027*	−0.287***	0.346***
0.2	−0.02	−0.131	−0.504***	0.008	0.044	0.065
0.3	0.039	0.145	−0.639***	0.003	−0.032	0.071***
0.4	0.074	0.421**	−0.503***	0.007	−0.104**	0.035
0.5	0.01	0.113	−0.484***	0.001	−0.165***	0.094***
0.6	−0.009	0.197	−0.094	0.006	−0.161***	0.093
0.7	0.003	0.512*	0.196	0.005	−0.122**	0.092*
0.8	0.098	0.391	0.49**	0.006	0.067	−0.059
0.9	0.065	0.214	−0.36	0.015	0.502***	−0.237*

Note(s): ***p < 0.01, **p < 0.05 and *p < 0.1; strong safe haven marked in italic

Source(s): Authors’ own creation

Table 5

Robustness test results: FGLS and GMM

	BAYC		MAYC		CP
	FGLS	GMM	FGLS	GMM	FGLS	GMM
SR	0.127	0.148	0.08	0.048	0.017	−0.007
FR	−0.199	−0.343	1.298***	2.06***	0.547***	0.878***
GR	0.276*	0.399**	−0.143	−0.104	−0.355**	−0.297**
GBR	−0.011	−0.022	−0.032	−0.07**	0.008	0.015
BTC	−0.078	−0.101*	−0.132	−0.131	0.048	0.051
ETH	0.153***	0.147***	−0.077	−0.116	0.054	0.034
No. of obs	1876	1848	1876	1848	1876	1848
Wald χ2	26.22	43.73	19.26	31.06	30.73	34.31
Prob > χ2	0.000	0.000	0.004	0.000	0.000	0.000

Note(s): ***p < 0.01, **p < 0.05 and *p < 0.1

Source(s): Authors’ own creation

Table A1

Indicators of SR, FR and GBR

Countries	SR	FR
The Philippines	PSEi Composite	PHP/USD
Thailand	SET Index	THB/USD
Malaysia	FTSE Malaysia KLCI	MYR/USD
Vietnam	VN Index	VND/USD
Brazil	Bovespa Index	BRL/USD
Hongkong	Hang Seng Index	HKD/USD
Peru	S&P Lima General Index	PEN/USD
Colombia	Colcap Index	COP/USD
South Africa	FTSE/JSE Top 40	ZAR/USD
Canada	S&P/TSX Composite Index	CAD/USD
Germany	DAX Index	EUR/USD
US	NASDAQ Composite Index
UK	FTSE 100 Index	GBP/USD
Japan	Nikkei 225 Index	JPY/USD

Note(s): 10-year government bonds are applied to all countries

Source(s): Created by authors

Notes

1. Finance Firms Seen Investing $1 Billion in Blockchain This Year, https://www.bloomberg.com/news/articles/2016-06-23/finance-firms-seen-investing-1-billion-in-blockchain-this-year (Accessed 2 June 2022).

2. Worldwide spending on blockchain solutions from 2017 to 2020, with forecasts for 2021 and 2024, https://www.statista.com/statistics/800426/worldwide-blockchain-solutions-spending/ (Accessed 6 June 2022).

3. Opensea statistics data, https://dune.com/rchen8/opensea (Accessed 21 May 2022).

4. NFT daily trading volume statistic, https://nonfungible.com/market-tracker (Accessed 23 May 2022).

5. Top three NFTs projects based on market capitalization, https://www.coingecko.com/en/nft

Declaration of conflicting interest: The authors declared no potential conflict of interest.

6. Bored Ape Yacht Club website, https://boredapeyachtclub.com/

7. CryptoPunks website, https://cryptopunks.app/

8. What Are CryptoPunks: A quick overview, https://www.blockchain-council.org/nft/cryptopunks/ (Accessed 30 May 2022).