Bachelor’s Thesis within the Bachelor’s Programme in Financial Economics Inflation and Asset Behavior: An Empirical Analysis of Bitcoin’s Hedging Potential Compared to Gold and Oil Abstract This study examines whether bitcoin exhibits inflation hedging characteristics and can be utilized by investors to preserve their capital from erosion caused by inflation. Furthermore, comparison with traditional inflation hedging assets such as gold and oil is made to determine whether bitcoin outperforms these assets. To accomplish this, two different linear regression models were constructed, where the first model focuses on the period between 2014-2024 and the second model focuses on the period 2019-2022. Moreover, this thesis is limited to U.S inflationary data. By dividing the thesis in two linear regression models, bitcoin's relationship with inflation can be analyzed both in the long-term and during times of heightened volatility in inflation, particularly before, during and after the COVID-19 pandemic. The results of the first model found no statistically significant relationship between bitcoin and inflation. However, the second model found a strong statistically negative relationship. Therefore, bitcoin may not serve as an inflation hedging asset. In comparison, gold and oil have been found from previous empirical studies to at least exhibit partial inflation hedging characteristics, implying a more suitable choice for investors worried about inflation. Victor Damberg Jacob Vähäkari Supervisor: Henrik Petri Spring Term 2025 15 ECTS Acknowledgements Special gratitude goes to our supervisor Dr. Henrik Petri for his guidance and support. There have been difficulties along the way, and while these were pointed out and discussed, his insights were valuable throughout the process. We also want to thank the opponents and fellow students for their feedback. Table of Contents 1. Introduction 1 1.1 Purpose 4 2. Theoretical framework 5 2.1 Inflation Hedging Theory 5 2.2 Safe-Haven Asset Theory 6 2.3 Stock-to-Flow Model 6 3. Previous Research 8 3.1 Hypothesis Development 11 4. Methodology 12 4.1 Research Design 12 4.2 Model Specification 12 4.3 Diagnostic Testing and Validity 14 4.4 Exclusion of Variables 16 4.5 Explanation of Variables 17 5. Data 19 5.1 Data Collection 19 5.2 Cleaning the Data 21 5.3 Descriptive Statistics 21 6. Empirical results 25 6.1 Bitcoin as an Inflation Hedge During 2014-2024 25 6.2 Bitcoin as an Inflation Hedge During the COVID-19 Pandemic 30 6.3 Gold and Oil as a Hedge Against Inflation 32 6.3.1 Gold as an Inflation Hedge 32 6.3.2 Oil as an Inflation Hedge 33 6.3.3 Conclusion 33 7. Discussion 34 7.1 Results 34 7.2 Limitations 37 8. Conclusion 39 9. Bibliography 40 1. Introduction In this initial chapter the reader is presented a broad general background regarding the topic set to be examined in the thesis. Incorporated in the background is a more niche discussion regarding the problems set to be investigated. Thereafter, justification and motivation regarding the problems is presented for the reader. Lastly, a research purpose is constructed where the research problems for this thesis is presented and concretized. Inflation continues to be a major concern that affects all individuals, investors, and institutions alike. The rise in prices results in the purchasing power decreasing over time, therefore, a growing interest in capital-preservation strategies has become a rising topic. For ages, commodities such as gold and oil were seen as strong inflation hedges, with great hedging capacity during times of economic instability (Baur and Lucey, 2010; Bodie, 1976). More recently, inflation protection discussions have opened up wide towards digital assets, especially bitcoin. Announced in 2009, bitcoin is a decentralized digital currency that works on its own, meaning without any need for control of a central authority or banks. Perhaps the most crucial factor in its attraction as a possible inflation hedge is that the total supply is capped at 21 million coins. Inflation may occur when constant amounts of traditional currency are made by central banks, contrasting with the limited supply of bitcoin (Weil, 2023; Kavinoky Cook LLP, 2024). Thus, bitcoin as "digital gold" is often claimed by proponents to help serve in the preservation of value. The discussion for bitcoin as an inflation hedge is built upon several arguments. Its decentralization and independence from government or central bank policies are regarded as a hedge opportunity against times of monetary uncertainty. Accessibility, borderlessness, and transparency, along with immutability underlining its blockchain mechanism, are two strong reasons for a significant interest in bitcoin as an asset. Such borderless characteristics make it an attractive option for those under financial hardships in their countries (Flores, 2024.; Jafar, 2025). Furthermore, with the rise in institutional demand for bitcoin, the legitimacy of bitcoin as a financial asset may get a boost. Corporations and investment firms are now relocating 1 portions of their portfolios to bitcoin, implying that an emerging consensus acknowledges bitcoin's role in diversified strategies to hedge against macroeconomic risks like inflation (Kavinoky Cook LLP, 2024). The advocates in favor of it even point out why bitcoin's pricing surged during the COVID-19 pandemic and after the stimulus, as inflation speculation began to gain traction. Yet, exactly how effective bitcoin can serve as a hedge against inflation remains uncertain. It is widely believed that the greatest counterargument against the performance of the digital asset lies in its excessive price volatility. Unlike gold or oil, where price movements are more stable and grounded in physical supply and demand, price movements of bitcoin are on the other hand influenced by speculation, market sentiment, and market news. The price fluctuations are often provoked by events such as regulatory crackdowns, exchange rate changes, and turns in macroeconomic conditions (Lyócsa et al. 2020; Fernandis 2024). Furthermore, while some studies indicate a positive correlation between bitcoin and inflation, it is important to find out if the correlation truly implies causality, or if this alignment arose mostly due to extrinsic factors. Such extrinsic factors may include macroeconomic trends, regulatory changes, and market sentiment, each of which has been shown to have considerable influence on bitcoin's price. In this complex backdrop, the debate in academia is yet to be settled, with some researchers giving credence to the fact that bitcoin can be relied upon to hedge against inflation, while others argue that its speculative nature and extreme volatility render it less suitable for this purpose. There are also opposing findings from empirical studies on both sides. Bitcoin may function as a hedge for inflation, particularly in high inflationary monetary environments (Choi & Shin 2022). However, the findings presented by Pinchuk (2023) denied this viewpoint through his research that bitcoin does not consistently possess hedging characteristics. Thus, in some situations reacts differently from inflation signals. In the same way, Smales (2024) writes about the possibility where bitcoin might show properties of being an inflation hedge under certain conditions. Nonetheless, regarding robustness, it is far a controversial take to state that bitcoin is a reliable hedge compared to traditional assets, such as gold. 2 Bitcoin, according to Taleb (2021), is perhaps best described as occupying a third space between a currency and a commodity. Hybridization makes it quite hard to categorize bitcoin fully since its characteristics resemble those of a number of asset classes. It is seen as a mode of exchange and unit of account in some communities, while on the other hand, it is more and more limited to speculative investment. This exclusive character means that the common financial models are unlikely to reflect how bitcoin adapts to inflationary pressures fully. Even more complexity is added to the discussion due to its historic valuation, speculation-oriented forecasting techniques like Stock-to-Flow, Metcalfe's Law, or market emotion analysis tend to dominate evaluations of bitcoin over the long run. Mixed opinions exist in the academic literature, these approaches have some pros and downfalls. Shelton (2024) also mentions that while offering insight, these models often have a few flaws when compared to real-world data. Especially during times when markets show peaks of stress or uncertainty. The narrative surrounding the inflation hedge bitcoin conspicuously does not really hold when subjected to gold comparisons. As reported by Shah (2025) and Royal (2024), "gold has been a store of wealth to humankind for thousands of years, with a time-tested track record across civilizations”. However, bitcoin is just over ten years old and has not been subject to an entire cycle of high and prolonged inflation in the most advanced economies. This time frame limits the ability to provide a strong foundation for concluding long-term effectiveness about its performance as a hedge. Despite its uncertainty, the interest in bitcoin as a hedge is still a hot topic. The younger generation, growing up in a technological world, are aware that bitcoin is more than just another asset. It is a new possibility to stay financially independent and resistant to vulnerability. Furthermore, the adoption of digital currencies in highly inflated countries such as Venezuela or Argentina may show how fiat currencies' trust may decrease over time. Still, in consideration, contradictory strands of thought and findings in the existing literature, this thesis seeks to explore whether in the longer time horizon, bitcoin can under certain circumstances behave as a hedge against inflation with respect to bitcoin returns. Comparing and contrasting the performances of bitcoin with other traditional hedges like gold and oil, it will finally examine the possible role of bitcoin as a reliable protection against inflation. 3 Deeper analysis may provide useful findings for individual investors, financial institutions, and policymakers looking for alternatives to preserving values. More importantly, it will also add insights to the growing discussion on positioning emerging digital assets as hedging material. 1.1 Purpose The purpose of this study is to investigate whether bitcoin is a reliable hedging tool against inflation, and thus expand on previous empirical research made regarding this topic. Previous empirical research is used as inspiration during this thesis, however, results from previous research may not be aligned with the results in this thesis. This thesis examines a relationship between bitcoin and inflation both during the 2014-2024 period as well as during a shorter time frame between 2019-2022. The shorter time frame represents the time before,during and after the COVID-19 pandemic. This differs from other empirical research conducted. Furthermore, this thesis will showcase results regarding the hedging ability of other commodities such as gold and oil, where previous research in this topic will be utilized. The purpose is to compare bitcoin with gold and oil’s capabilities as a hedge against inflation rather than examining if gold and oil are hedges themselves, since there is comprehensive research done regarding this topic already. Since gold and oil are already considered traditional hedging assets, comparison with bitcoin characteristics is vital in order to establish whether there is any relationship between bitcoin and inflation. Based on this, the main research question is defined as follows: Can bitcoin serve as a reliable hedge against inflation compared to gold and oil during the period 2014–2024 and the COVID-19 pandemic? 4 2. Theoretical framework In this chapter the reader is presented with elementary information of economic theories which the thesis utilizes as a majority foundation for concepts that are used in the following chapters. This chapter aims to create clarity and thus aid the reader’s understanding for the following chapters. Inflation hedging theory, safe-haven asset theory and the stock-to-flow model are the theories described and accounted for and presented in the same order declared here. 2.1 Inflation Hedging Theory The inflation hedging theory is a concept that clarifies if an asset has the capability to increase or maintain its own real value throughout times with increasing inflation. If it does, then that asset can be used to hedge inflation. With increasing inflation, the price of goods and services rise, and results in a weaker purchasing power of money in the economy. However, asset returns which have a positive correlation with inflation allow investors to protect their investments from inflation and this may also create growth of their wealth when considering real term valuation. Therefore, positively correlated asset returns with inflation can be considered an effective inflation hedge. Furthermore, in order to classify an asset as an inflation hedge, the positive correlation between real returns and inflation must be stable over a longer time period. (Bodie, 1976). According to Bodie (1976), there are other characteristics than those mentioned in the preceding paragraph that can be used to identify if an asset is a suitable inflation hedge, namely intrinsic value and supply constraints. An effective inflation hedge should not be purely affected by financial instruments and market sentiment. The asset should have its own intrinsic value, a fundamental value that is not affected by financial instruments and market sentiment. Commodities such as gold and oil are typical inflation hedges, who also fulfill the last criteria about supply constraints. This criteria states that assets that have limited supply tend to be stronger inflation hedges since they perform better due to the values of these assets being less sensitive to monetary dilution. The distinction between real and nominal returns are essential to understand the effectiveness of an inflation hedge asset. An asset with increases in nominal return while real return does 5 not keep the same pace as inflation can not be classified as an effective inflation hedge since the purchasing power is not preserved. (Bodie, 1976). 2.2 Safe-Haven Asset Theory The definition of a safe haven is “A safe haven is defined as an asset that is uncorrelated or negatively correlated with another asset or portfolio in times of market stress or turmoil”.(Baur and Lucey, 2010). A safe haven asset is therefore an asset that can be both positively and negatively correlated during normal market conditions while it can be negative in extreme market conditions. However, it is worth noticing that a safe haven asset can exhibit negative correlation to other assets during extreme market conditions while still having a positive correlation towards inflation, e.g. during time of stagflation. Therefore, if the safe haven asset manages to have a negative correlation with inflation during extreme market conditions, the asset will lose value while inflation rises, indicating it is not a suitable hedging tool. Furthermore, it is essential to determine the inflation condition during periods with negative correlation. More specifically between the asset and the market during extreme conditions, in order to be able to classify the asset as a safe-haven asset or not. This can be connected to the first theory, inflation hedging theory (Bodie, 1976), where an effective inflation hedge asset should have positive correlation over time with inflation. If this condition is met, the safe-haven asset theory can be applied to this study in order to showcase whether or not bitcoin, compared to gold and oil, are suitable hedging assets during normal and extreme market conditions or both. 2.3 Stock-to-Flow Model The Stock-to-flow model has traditionally been used to evaluate the price of commodities such as gold. In 2019, a pseudonymous Dutch blogger proposed the use of the Stock-to-flow model in order to show how the scarcity and price of bitcoin is related, which is discussed further by Shelton (2024). To use the model, a ratio between the total existing supply of bitcoin and its newly minted supply shows the scarcity of bitcoin. Therefore, bitcoin is classified as a scarce commodity according to this model. The relation between the supply of bitcoin and its newly minted supply shows how valuable the bitcoin is at that moment, where a higher level of scarcity implies a higher valuation of the bitcoin. This is true when you assume a constant or increasing demand for bitcoin. 6 Another assumption by this model is that events known as halvings will occur every four years, where miners of bitcoin will receive less amount of bitcoin’s for validating transactions. According to (CoinMarketCap, n.d.), every halving occurs after 210 000 blocks are mined, where each mined block results in a reward for the miner, namely receiving bitcoin’s. The amount of bitcoin the miners receive is therefore reduced by half of the current amount every fourth year, which will result in less new bitcoin’s coming into circulation. This also results in a higher valuation of the bitcoin since the newly minted supply will decrease. (Shelton, 2024). 𝑆2𝐹 = 𝑇𝑜𝑡𝑎𝑙 𝐶𝑖𝑟𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑆𝑢𝑝𝑝𝑙𝑦 𝑜𝑓 𝐵𝑖𝑡𝑐𝑜𝑖𝑛 𝑁𝑒𝑤𝑙𝑦 𝑀𝑖𝑛𝑡𝑒𝑑 𝑆𝑢𝑝𝑝𝑙𝑦 𝑜𝑓 𝐵𝑖𝑡𝑐𝑜𝑖𝑛×210 000 This is the formula used to calculate bitcoin's scarcity taking into account the halving events. Since this formula measures the scarcity over a 4 year period, it will be adjusted to a monthly rate to fit this study. Therefore 210 000 will be divided by four to get the yearly rate of blocks mined, and then divided by 12 to get the monthly rate of blocks mined. The new formula becomes: 𝑆2𝐹 = 𝑇𝑜𝑡𝑎𝑙 𝐶𝑖𝑟𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑆𝑢𝑝𝑝𝑙𝑦 𝑜𝑓 𝐵𝑖𝑡𝑐𝑜𝑖𝑛𝑁𝑒𝑤𝑙𝑦 𝑀𝑖𝑛𝑡𝑒𝑑 𝑆𝑢𝑝𝑝𝑙𝑦 𝑜𝑓 𝐵𝑖𝑡𝑐𝑜𝑖𝑛×4375 When the stock-to-flow model holds in the data, it may provide solid justification for using bitcoin as a hedge against inflation because the model does not include other macroeconomic variables. According to this model, the value of bitcoin should only be able to increase unless new minted supply increases. Since a higher scarcity increases the price of bitcoin, it will also rise during times with high inflation, indicating a suitable inflation hedging asset. Therefore, empirical testing of the models predictions will be vital in order to examine the actual relationship between inflation and bitcoin. 7 3. Previous Research Firstly, in this chapter the reader is presented with a summation of previous research regarding similar topics examined in this thesis. The previous research is fundamental for the formation of the thesis and further provided justification for theoretical standpoints utilized. Secondly, the reader is presented with a hypothesis development which details the thought process behind the hypothesis used and influenced by the previous research. Smales (2024) covers the topic of using cryptocurrencies, mainly bitcoin, as an inflation hedge and whether or not it is an effective hedging asset. He mentions the rising demand and popularity of digital assets, particularly during times of uncertainty in the market. Smales’ research examines the value preservation of purchasing power by using cryptocurrencies as a tool of hedging inflation in the same way as traditional assets such as gold are used in inflation hedging. Furthermore, to examine this relationship, Smales makes four key assumptions as stated below. Inflation Hedging Theory: During times with high inflation, investors want assets that retain their value and opt for other assets than fiat currencies, whose purchasing power decreases during times with high inflation. This aligns with the argument that bitcoin, like gold, could serve as a store of value in times of inflation. Moreover, the assumption of bitcoin as “Digital Gold” builds on this idea, emphasizing that the scarcity of bitcoin is a similar characteristic that gold has, and together with decentralized properties, bitcoin may act as a potential hedging asset. Additionally, this perspective aligns with the argument that bitcoin's fixed supply could make it a valuable store of value. Both Smales (2024) and Choi & Shin (2022) emphasize the pivotal role of market efficiency when considering whether an asset can be said to actually function as a hedge. Market efficiency and speculation is another set of assumptions that Smales makes when suggesting that bitcoin, like many other cryptocurrencies, carries a high degree of volatility. The volatility might be basically related to speculative demand, which sets the price of bitcoin, and not to economic fundamentals. Similarly, Choi & Shin (2022) highlights the importance of market efficiency to any asset that attempts to act as a stable store of value. They define that the price volatility of bitcoin on many occasions reflects speculative interests, as opposed to economic fundamentals, and in this way being used as a solid frontier against inflation. 8 Macroeconomic influence is the last assumption considered by Smales, where he states that any price trigger, including confounding shocks of inflation, would affect asset price movements. Therefore, if bitcoin were an actual hedge, the price would need to follow inflationary pressures. Smales (2024) and Choi & Shin (2022) find it difficult to argue that this influence actually holds. Given that bitcoin's extreme volatility often inhibits this assumption, indicating that price movements are unpredictable and rarely correlate with inflation data as would be expected of a conventional hedge. Building on this, Smales implemented three models to examine the relationship between cryptocurrencies and inflation: time-series analysis, linear regression models, and comparative analysis during 2011-2022. Time-series analysis was included to find responses to the historical price change in bitcoin due to inflation. The linear regression models were included to test the correlation between inflation rates and returns from bitcoin and other cryptocurrencies. The comparative analysis was used to compare the performance between common inflation hedges such as gold and bitcoin’s performance. However, Smales found no strong consistent relationship between inflation and bitcoin returns, mainly due to the high volatility of bitcoin. Although during shorter periods, bitcoin seemed to have been efficient as a hedge against inflation, these observations further clarified that it was mostly random and hard to predict beforehand when bitcoin could have been used as an efficient inflation hedging asset. Likewise, Choi & Shin (2022) examined the relationship between bitcoin and inflation, where inflation is measured by the consumer price index (CPI). Interest rates and money supply growth are also some economic variables examined to find a relationship with bitcoin. This thesis also includes assumptions in order to examine the relationship between bitcoin and inflation. These are similar to the ones mentioned in the previous article written by Smales. Some of the assumptions used in the analysis are market efficiency and bitcoin as a store of value. The market efficiency assumption indicates that the price of bitcoin should react efficiently to macroeconomic indicators. Furthermore, the inflation expectations should be reflected in eventual price movements for bitcoin. Bitcoin being seen as a store of value implies that bitcoin has a fixed supply, which may indicate that bitcoin could serve as an efficient inflation hedge along the lines of gold, which also has a fixed supply. However, the authors of this study also used regression analysis, as well as correlation analysis, and made comparisons with other assets in order to establish a relationship between bitcoin and 9 inflation. Despite this effort, the authors of this work found that bitcoin is very volatile and therefore cannot act as a reliable inflation hedge consistently. The authors also conclude that bitcoin should be seen as a speculative asset with high risk rather than a suitable asset to hedge inflation, since it does not provide reliable protection against inflation. Pinchuk (2023) investigates if bitcoin can be considered a trusted hedge towards inflation by observing its reaction to macroeconomic happenings, especially when there are surprises to inflation data. The high-frequency event-based methodology measures the price response of bitcoin to inflation movements. The main finding is that bitcoin tends to react negatively, more precisely, one standard deviation positive inflation surprise will convert into a decline in bitcoin’s price by 24 basis points. This reaction pattern is indicative of the finding that bitcoin behaves more like a risky asset, rather than hedging. Interestingly, Pinchuk also ascertains if this reaction is connected to exposure to interest rates, but no evidence is found to support his ideas. Instead, the change in consumption-saving behaviors due to inflation expectations is found to affect the price of bitcoin. Moreover, other cryptocurrencies examined in this study did not show substantial responsiveness to inflation news, highlighting that bitcoin behaves uniquely compared to other cryptocurrencies. Finally, Taleb (2021) adopts a broad macroeconomic perspective to determine whether bitcoin behaves like currency, a hedge, or a-haven asset. The study dives into how bitcoin responds to inflation, exchange rates, and financial market fluctuations. Through econometric modeling and regression analysis, the results showed that it is extremely reactive to fragility in the financial system and sentiment in global markets. This suggests that bitcoin is more of a speculative asset than a reliable hedge, aligning with the findings of Pinchuk and Choi & Shin. Together, these studies indicate that while bitcoin has certain characteristics that could theoretically support its role as an inflation hedge, its extreme volatility and inconsistent behavior in response to macroeconomic factors significantly limit its reliability as a protective asset. 10 3.1 Hypothesis Development Previous studies have reached contradictory results about this inflation-hedging property posed by bitcoin. Some researchers argued that bitcoin's fixed supply and decentralization are key features supporting bitcoin being an inflation hedge. On the other hand, there are those who highlight the extreme volatility of bitcoin prices and their speculative nature, making bitcoin unreliable. The present study thus intends to test empirically if, over time, and especially during periods of economic stress, bitcoin has had statistically significant relationships with key inflation indicators. Given this background, the following hypothesis is proposed: 𝐻 : There is no statistically significant relationship between bitcoin returns and inflation 0 indicators. 11 4. Methodology This chapter presents the reader with the methodological approach used to investigate bitcoin's inflation hedging characteristics. Furthermore, regression model(s) utilized are presented and abbreviations are explained in detail. Moreover, to ensure the model(s) validity, diagnostic testing methods are presented. Lastly, limitations are set where exclusion of alternative variables are discussed and detailed descriptions of the variables included in the model(s) are given. 4.1 Research Design This study will have a quantitative approach of research to examine to which extent bitcoin can serve as a hedge against inflation. The study pursues an analysis of the relationship of bitcoin returns with inflation data over a long time period, to be able to provide meaningful insights. The scope of empirical analysis is going to include historical data on returns of bitcoin, the Consumer Price Index (CPI) as the primary measure of inflation, and the price shifts of traditional inflation hedges like gold and oil. Data will be collected from trusted databases and sources. This includes CoinMarketCap, the Federal Reserve Bank of St.Louis, U.S Bureau of Labor Statistics, and other alike financial data providers. 4.2 Model Specification The independent variables in the regression are chosen based on theory and empirical evidence associating them with inflation dynamics and bitcoin price behavior. Core CPI captures basic inflation trends, breakeven inflation rates stands for market inflation expectations, VIX index is a measure for market stress and the stock-to-flow ratio measures bitcoin supply scarcity. Together these cover both macro-based and asset-specific factors that ought to affect bitcoin returns. Statistical tools such as the Pearson correlation and two different time-series OLS regression analyses will be conducted to determine the strength and stability of the relationship between movements of inflation and different asset returns on a 5% significance level. Special emphasis will be made on periods of very elevated inflation, economic stress, and extreme monetary responses but also during normal market and inflation conditions. 12 Pearson Correlation: Σ[(𝑥 −?̄?)(𝑦 −ȳ)] 𝑟 = 𝑖 𝑖 𝑥𝑦 Σ(𝑥 −?̄?)2× Σ(𝑦 −ȳ)2] 𝑖 𝑖 Multiple regression model: during 2014-2024. 𝑙𝑛(𝐵𝑇𝐶 ) = α + β₁·Δ𝐶𝑃𝐼 (𝐶𝑜𝑟𝑒) + β₂·Δ𝑇10𝑌𝐼𝐸𝑀 + β₃·Δ𝑉𝐼𝑋 ² + β₄·𝑙𝑛(Δ𝑆2𝐹 ) + ε 𝑡 𝑡 𝑡 𝑡 𝑡 𝑡 Multiple regression model: during 2019-2022. 𝑙𝑛(𝐵𝑇𝐶 ) = α + β₁·Δ𝐶𝑃𝐼 (𝐶𝑜𝑟𝑒) + β₂·Δ𝑇5𝑌𝐼𝐸𝑀 + β₃·Δ𝑉𝐼𝑋 ² + β₄·𝑙𝑛(Δ𝑆2𝐹 ) + ε 𝑡 𝑡 𝑡 𝑡 𝑡 𝑡 Where: - 𝑙𝑛(𝐵𝑇𝐶 ): Logarithm of bitcoin returns at time t 𝑡 - Δ𝐶𝑃𝐼 (𝐶𝑜𝑟𝑒): Monthly change in core inflation (Consumer Price Index) 𝑡 - Δ𝑇10𝑌𝐼𝐸𝑀 : Monthly change in 10-year inflation expectations 𝑡 - Δ𝑇5𝑌𝐼𝐸𝑀 : Monthly change in 5-year inflation expectations 𝑡 - Δ𝑉𝐼𝑋2: Monthly change in squared volatility index 𝑡 - 𝑙𝑛/(Δ𝑆2𝐹 ) : Natural log of monthly change in Stock-to-Flow ratio 𝑡 - ε : Error term 𝑡 As seen in the formulas above, both regressions contain the same variables except for the breakeven inflation rate. The reasoning behind this is that the 10-year breakeven inflation rate is supposed to capture the long term (10 years) expected inflation in the first regression model, where bitcoin and inflations relationship is tested during 2014-2024. Using the short term breakeven inflation rate (5 years) in the first regression model may be misleading and resulting in the long term relationship between bitcoin and inflation not being captured. 13 Vice versa can be said for the second regression model, where the aim is to capture bitcoin and inflations short-term relationship in the same period. Using the short term breakeven inflation rate is therefore more suitable to use in the second regression. Lastly, to calculate the marginal effect for the variables, two different approaches are used. Firstly the log transformed stock-to-flow ratio is calculated straightforward, where a one percent increase in the stock-to-flow ratio changes the log transformed bitcoin returns with (LN S2F) percent. Secondly, the other variables are stated in monthly percentage changes in decimal form. Since bitcoin returns are log transformed, the following formula is used to calculate the marginal effect when increasing the independent variable with a percentage point. (𝑒β*∆𝑋 − 1) * 100 4.3 Diagnostic Testing and Validity To ensure that the output from the regression is not disturbed by multicollinearity among the independent variables, a Variance Inflation Factor (VIF) will be utilized. The variables that exceed a value of 5 will be carefully reviewed or adjusted. 𝑉𝐼𝐹 = 1 𝑗 (1−𝑅2) 𝑗 Where: - 𝑉𝐼𝐹 : Variance Inflation Factor for variable j. 𝑗 - 𝑅2: Coefficient of determination from regressing variable j on all other independent 𝑗 variables. 14 Besides multicollinearity testing, some diagnostic tests will be made in an attempt to validate the statistical assumptions underlying the regression model. An ADF test will be conducted to confirm the assumption of the time series data composition, to guarantee that spurious regression caused by non-stationary variables does not occur. A p-value below 0.05 indicates stationarity. Δ𝑦 = 𝑎 + β + γ𝑦 + δ Δ𝑦 +... + δ Δ𝑦 + ε 𝑡 𝑡 𝑡−1 𝑖 𝑡−𝑖 𝑝 𝑡−𝑝 𝑡 Where: - Δ𝑦 : First difference of the time series at time t 𝑡 - α: Constant/intercept - β : Linear time trend 𝑡 - γ𝑦 : Lagged level of the dependent variable 𝑡−1 - δ Δ𝑦 : Lagged differences of the dependent variable (for i = 1 to p) 𝑖 𝑡−𝑖 - ε : Error term at time t 𝑡 A Durbin-Watson statistic will also be utilized to validate for autocorrelation in the residuals where values close to two means independence of observations and values much higher/lower indicate presence of serial correlation, thus biasing the coefficient estimates. The Durbin-Watson statistic is being utilized to find autocorrelation of residuals that would otherwise create bias against regression coefficients and inference. Σ(𝑒 −𝑒2 ) 𝐷𝑊= 𝑡 𝑡−1 Σ𝑒2 𝑡 Where: - 𝑒 : Residual at time t 𝑡 - 𝑒 : Residual at time t−1 𝑡−1 - Σ: Summation over the sample period 15 Lastly, a Breusch-Pagan test will be made to assess any presence of heteroskedasticity that is non-constant variance of the error terms. If heteroskedasticity exists, robust standard errors may be estimated to make valid inferences. The test relies on two different versions: the Lagrange Multiplier (LM) test and an F-statistic. A p-value>0.05 indicates homoskedasticity, while on the other hand a p-value lower than 0.05 indicates that heteroskedasticity is present and must be handled. The Breusch-Pagan test helps identify heteroskedasticity, ensuring the validity of standard errors and the overall fit of the model. 𝐵𝑃 = 𝑛 × 𝑅2 Where: - 𝑛: Number of observations - 𝑅2: Coefficient of determination from the auxiliary regression of squared residuals These tests will bolster the regression results by ascertaining the validity of the underlying econometric assumptions. Moreover, this study will examine the hedging capability of bitcoin not just from correlation, but also through volatility, and reactiveness to external shocks. The analysis also adjusts for significant global macroeconomic events. Additionally, the study makes an in depth attempt to compare the performance of bitcoin alongside gold and oil, to further distinguish how well bitcoin performs as a hedge in comparison to traditional ways of hedging inflation. 4.4 Exclusion of Variables Lags of the independent variables, change in Core CPI, ln(S2F), and change in VIX, are mostly introduced to take any delayed reactions into account that may have occurred in the markets. Lags of 1, 2, and 3 months in this study were tested. In all instances, these lagged versions reduced the model-fit and explanatory power of the equations produced, thus eliminating all the variables with lags while retaining only the contemporaneous versions. Other metrics were tested as the main inflation measure, such as Trimmed Mean PCE, Median CPI, and headline CPI. However, these measures were excluded because of poor model fit impact. Sharpe ratio was considered but excluded because it lacks strong theoretical 16 backing for inflation hedging in this context. In the end, Core CPI was selected due to the fact that it removes a large amount of volatility coming from more volatile components like food and energy, and thus can be seen as a more stable and policy-relevant measure of inflation for this thesis. 4.5 Explanation of Variables Natural logarithm of bitcoin returns (𝑙𝑛(𝐵𝑇𝐶 )) 𝑡 The dependent variable in the model is the natural logarithm of monthly returns on bitcoin. However, instead of absolute price changes, logarithmic returns are rather used to stabilize the variance, control skewness and enable interpretation in percentage terms. Actually, the same transformation is often used for all of financial econometrics. A precise model is produced as this decreases the influence of outliers. (Rodriguez & Colombo, 2025; Shelton, 2024). The goal of the study is to analyze to which extent bitcoin has become a reliable hedge for inflation. This variable will measure how bitcoin reacts with respect to include macroeconomic factors. Core Inflation Change (Δ𝐶𝑃𝐼 ) 𝑡 Measuring core CPI cuts real inflation from the influence of extremely volatile variables such as food and energy prices. (Rodriguez & Colombo, 2025; Smales, 2024). The monthly change has been taken into account in the model to analyze whether bitcoin responds to relative inflation reality. Core inflation is the most significant benchmark for both monetary policy and economic stability, thus making it relevant for studying the hedging effect against inflation. 10-Year Breakeven Inflation Rate Change (Δ𝑇10𝑌𝐼𝐸𝑀 ) 𝑡 This variable captures the expectations the market has of long-term inflation, given by the difference between governmental nominal bonds and inflation-protected bonds. It is a future oriented measure and also includes its monthly change in the model for regression in order to analyze if anticipated inflation would change bitcoin's return pattern. If bitcoin is to act as a hedge, it will be affected by both current and expected inflation. (Wissmann, 2022; Weil, 2023). 17 5-Year Breakeven Inflation Change (Δ𝑇5𝑌𝐼𝐸𝑀 ) 𝑡 This variable captures the expectations the market has of medium-term inflation. In the same way as the 10-year breakeven inflation rate, the 5-year breakeven inflation rate is given by the difference between governmental nominal bonds and inflation-protected bonds and is a future oriented measure. Given that this variable measures monthly changes in the 5-year breakeven inflation rate, both current and anticipated inflation can be analyzed to identify if bitcoin’s return pattern changes. If bitcoin is expected to be a hedge, it will react to both current and the expected inflation. (Wissmann, 2022; Weil, 2023). Squared Change in Volatility Index (Δ𝑉𝐼𝑋2) 𝑡 The VIX index shows the market in the short-term as it is considered to act as a "fear index" , while at the same time still capturing changes in VIX. (Köse et al., 2024; Taleb, 2021). The model visualizes the effect of periods of economic stress and uncertainty in the market by involving this squared monthly difference into the model. Squaring the change favors sharp rises in volatility, which is important to identify if bitcoin poses as a safe haven or reacts to market turmoil differently. Natural logarithm of Change in Stock-to-Flow Ratio (𝑙𝑛Δ𝑆2𝐹 ) 𝑡 This measure was performed especially for bitcoin on a micro-cryptocurrency scale. Stock-to-flow is defined as the relationship of the total existing supply of bitcoin’s to the annual production of new ones. It is commonly used in the crypto community to understand and even predict the price of bitcoin, relative to its scarcity. (Morillon & Chacon, 2022; Shelton, 2024). The model shows how marginal changes in the anticipated scarcity of bitcoin, calculated by natural logarithm of the monthly change in the stock-to-flow model (S2F) ratio, affects its price with respect to inflation. 18 5. Data In this chapter the reader is presented with the data collection process. Furthermore, justifications and motivations regarding where and how the data collection was gathered are being discussed. Moreover, removal of excess data points retrieved and retention of outliers are being considered to maintain academic integrity. At last, descriptive statistics are presented where graphs containing bitcoin and inflation trends, a correlation matrix and a table consisting of transformations/calculations utilized to each variable included in the model(s) are visible. 5.1 Data Collection To test the relationship between bitcoin and inflation, bitcoin closing price data from CoinMarketCap was assembled. Considering bitcoin is a cryptocurrency and not an asset such as a stock, that can be included in any index, only the monthly bitcoin closing prices were collected. CoinMarketCap is a prominent website for cryptocurrencies and provided historical closing prices between the first of January 2014 to the first of January 2025, resulting in 133 different data points. The time period was selected to capture a decade of bitcoin activity during low and high times of inflation. It also includes the COVID-19 pandemic and the subsequent recovery. The time frame, therefore, facilitates a thorough assessment of bitcoin's behavior amid a variety of macroeconomic conditions. Furthermore, this depth of data points was essential in order to make significant statistical inference. Therefore, CoinMarketCap was the most suitable choice of database in order to construct our dataset since other cryptocurrency websites lacked the depth and accuracy of bitcoin closing price data. To find a relationship between bitcoin and inflation, an inflation measure was gathered. Core cpi inflation was chosen as the measure and the data was collected from the U.S Bureau of Labor Statistics Data database over monthly core cpi for the U.S during 2014-2024. 133 data points were collected to match the time frame of the other variables. The U.S Bureau of Labor Statistics is an official website of the U.S government and is one of the most prominent sources of inflation statistics for the U.S. 19 The dataset used also consists of variables such as Cboe´s implied volatility index for the S&P 500 options, the 10-Year breakeven inflation rate and the 5-year breakeven inflation rate. The implied volatility index for the S&P 500 options were gathered from Cboe´s own database, where monthly observations throughout the entire timespan were available. Both the 10-Year and the 5-Year breakeven inflation bonds were retrieved from the Federal Reserve Bank of St.Louis's database, where extensive monthly yields for the bonds were available. For the Cboe’s implied volatility index, the 10-Year breakeven inflation bond and the the 5-Year breakeven inflation bond 133 data points were collected each, resulting in monthly data for all of the variables to match the bitcoin closing price data. Lastly, to measure the stock-to-flow models impact on bitcoin valuation, data regarding total circulating supply of bitcoin and newly minted supply of bitcoin each year was collected. This was gathered from another prominent cryptocurrency website called bitbo. Together with CoinMarketCap, bitbo has one of the most extensive historical data regarding bitcoin and has data about total circulating supply of bitcoin and newly minted supply on a yearly basis. Also, both bitbo and CoinMarketCap provided data regarding halving events for bitcoin miners, which changes the newly minted supply, affecting the value of bitcoin. After compiling this data, calculations could be made to determine a stock-to-flow ratio each month. Again, the sample dataset consists of 133 monthly observations of the stock-to-flow ratio, matching the time frame set for the period 2014-2024. The reason for including 133 data points for each variable in the dataset was to assure significant statistical inference as well as making sure that the time frame included a period with heightened market uncertainty and higher inflation, e.g. before, during and after the COVID-19 Pandemic. A second OLS regression was also made in order to find bitcoin's relationship with inflation in a shorter time frame. The time frame investigated is the period 2019-2022. This essentially means that the sample dataset was reduced to 49 observations to test the short term relationship between bitcoin and inflation before, during and after the COVID-19 pandemic. In this regression the same variables were used, except for the 10-year breakeven inflation rate, which was changed to the 5-year breakeven inflation rate instead. 20 5.2 Cleaning the Data Some cleaning of the dataset has been made. Consider for example the stock-to-flow ratio and the data needed to calculate the ratio (e.g. total supply and newly minted supply of bitcoin). In the database where the data was assembled, sorting the data to fit the time frame in this thesis was not possible. Therefore, all the available data was gathered and resulted in excess data points outside of the chosen time frame in the sample data. This led to the removal of both the data points outside the chosen time frames and also the removal of daily observations, since this thesis only considers monthly observations. No other data that was collected was adjusted or cleaned in any way, mostly to preserve the reliability of the research conducted in this thesis but also to keep the original data points as they were even if outliers occurred. Keeping the outliers was crucial to examine whether or not bitcoin can act as a hedge both during normal market conditions but also during times with market turmoil such as during the COVID-19 pandemic. 5.3 Descriptive Statistics Figure 1 Shows the monthly log closing prices for bitcoin during 2014-2024. Figure 1. Closing log prices of bitcoin between the first of january 2014 and the first of january 2025 As seen in figure 1, bitcoin has had a long term upward looking trend regarding its price. Although, in some periods, mainly between 2018-2019 and 2022-2023, the price of bitcoin showcases a downward looking trend. During these time intervals no financial crisis nor market turmoil was affecting the price. However, it is possible that speculative cycles or other macroeconomic factors affected the price during these downward trends. 21 In Figure 2, the log of bitcoin’s monthly returns between 2014-2024 is visualized. Figure 2. Monthly natural logarithm of bitcoin returns between the 1st of January 2014-the 1st of January 2025 As seen in figure 2, bitcoin returns seem to be very volatile during the whole sample period. However, the returns also seem to become less volatile in recent years. In the beginning of 2021, a big drop in the returns can be spotted, but afterwards the returns seem to stabilize compared to the years prior. During 2020-2022, the COVID-19 pandemic caused economic disruptions globally, however the bitcoin returns seems to have stabilized somewhat after the big drop in the beginning of 2021. Figure 3 shows the monthly absolute levels of core cpi data for the U.S during 2014-2024. Figure 3. Absolute monthly levels of core cpi data between the 1st of January 2014- 1st of January 2025 in the U.S. (In the regressions monthly changes of core cpi data is used and not the absolute level). In figure 3 a clear long term positive trend of the core cpi can be seen. During 2014-2020 the positive trend seems to be stable and increase roughly as much each year. In 2020 the core cpi 22 increases with a decreasing rate compared to the years prior, indicating the start of the COVID-19 pandemic. From 2021 and onwards the core cpi seems to increase with an increasing rate again, increasing at a more rapid pace than during 2014-2020. This increasing increase between 2021-2024 may be explained from the U.S economy recovering from the COVID-19 pandemic, which created disruptions in the U.S economy during 2020. Figure 4 shows a correlation matrix between all variables included in all final regressions. Figure 4. A correlation matrix between all variables included in the final regressions used in the thesis. As seen in figure 4, all variables showcase a low correlation, except for the T10YIEM Change and T5YIEM Change variables. However, the T10YIEM and T5YIEM variables are not included at the same time in any of the regressions, resulting in the regressions not having an issue with high correlation of the variables included. 23 Table 1. Variable names, description and relevant transformations/calculations used. Variable Variable description Transformation/calculation LN Returns Natural Logarithm of bitcoin 𝐿𝑁 (𝐵𝑇𝐶 ) − 𝐿𝑁 (𝐵𝑇𝐶 ) 𝑡 𝑡−1 monthly returns at time t. CPI (Core) Monthly change in core inflation (𝐶𝑃𝐼 (𝐶𝑜𝑟𝑒) ) − (𝐶𝑃𝐼 (𝐶𝑜𝑟𝑒) )𝑡 𝑡−1 Change at time t. (𝐶𝑃𝐼 (𝐶𝑜𝑟𝑒) )𝑡−1 T10YIEM Monthly change in 10-year (𝑇10𝑌𝐼𝐸𝑀 ) − (𝑇10𝑌𝐼𝐸𝑀 )𝑡 𝑡−1 Change breakeven inflation rate at time t. (𝑇10𝑌𝐼𝐸𝑀 )𝑡−1 T5YIEM Monthly change in 5-year (𝑇5𝑌𝐼𝐸𝑀 )− (𝑇5𝑌𝐼𝐸𝑀 )𝑡 𝑡−1 Change breakeven inflation rate at time t. (𝑇5𝑌𝐼𝐸𝑀 )𝑡−1 VIX Monthly change in squared First step: Calculate the change in Change^2 volatility index at time t. VIX. (𝑉𝐼𝑋 ) − (𝑉𝐼𝑋 ) 𝑡 𝑡−1(𝑉𝐼𝑋 ) 𝑡−1 Second step: Square the change in VIX. (𝑉𝐼𝑋 𝐶ℎ𝑎𝑛𝑔𝑒 )2 𝑡 LNS S2F Natural logarithm of monthly First step: Calculating the Change change in Stock-to-Flow ratio at Stock-to-Flow ratio. time t. 𝑇𝑜𝑡𝑎𝑙 𝐶𝑖𝑟𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑆𝑢𝑝𝑝𝑙𝑦 𝑜𝑓 𝐵𝑖𝑡𝑐𝑜𝑖𝑛𝑡 𝐴𝑛𝑛𝑢𝑎𝑙 𝑃𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑜𝑛 𝑜𝑓 𝐵𝑖𝑡𝑐𝑜𝑖𝑛 × 4375 𝑡 Second step: Log transforming the Stock-to- Flow ratio and then calculating the monthly change. 𝐿𝑁 (𝑆2𝐹 ) − 𝐿𝑁 (𝑆2𝐹 ) 𝑡 𝑡−1 Table 1. This table details the variables used in the regressions together with short descriptions of each variable and showing the transformations/ calculations made before using each variable. 24 6. Empirical results In this chapter the reader is first introduced to multiple different regression models aimed to show the academic progression towards the final model presented in subsection 6.1. To arrive at the final model, diagnostic testing is presented to exhibit arguments for transforming the variables in the model. The initial subsection covers the period 2014-2024. The second subsection utilized the same regression model arrived at in subsection 6.1 with a small change in breakeven inflation rate. This model covered the period 2019-2022. Lastly, empirical evidence from past studies regarding gold and oil as inflation-hedging assets are presented. 6.1 Bitcoin as an Inflation Hedge During 2014-2024 Table 2 shows the first OLS regression result for the time period 2014-2024. This was the first regression conducted and all variables except bitcoin returns are in its absolute index form, e.g. not transformed. The bitcoin returns are transformed into a natural logarithm. This was done to establish a baseline. The log transformed bitcoin returns is used as the dependent variable. Variable Coefficient Std.Error T-stat Significance Intercept 0.0270 0.0446 0.6052 CPI (Core) 0.0001 0.0002 0.2095 T10YIEM -0.0169 0.0074 -2.2800 ** VIX -0.0261 0.0324 -0.8050 S2F ratio 0 0 0.6627 𝑅2 0.042 Observations 133 Table 2. All values are rounded to 4 decimal points. ***,**.* indicate significance at 1%, 5% and 10% respectively. As seen in table 2, using the absolute index values of the independent variables results in a 𝑅2 value of 0.042, where all variables are statistically insignificant except for the 10-year 25 breakeven inflation rate, which is statistically significant at both 10% and 5% significance level. This results in the absolute level for the 10-year breakeven inflation rate seems to have a small negative effect on bitcoin returns during the 2014-2024 sample period with a coefficient of -0.0169. However, in table 3, results from tests regarding stationarity, multicollinearity, autocorrelation and heteroskedasticity are shown. Variable/element Stationarity Multicollinearity Autocorrelation Heteroskedacity (ADF) (VIF) (DW) (BP) LN Returns Stationary - - - CPI (Core) Non-Stationary 5.7320 - - T10YIEM Non-Stationary 1.5751 - - VIX Stationary 1.09 - - S2F-Ratio Non-Stationary 5.2698 - - Model Residuals - - 1.7988 Present Table 3. Results from tests regarding stationarity, multicollinearity, autocorrelation and heteroskedasticity in both the variables and the model's residuals. All values are rounded to 4 decimal points. From table 3, it is clear that the initial model faces difficulties in the form of non-stationarity for the variables CPI (Core), T10YIEM and S2F-ratio, multicollinearity for the variables CPI (Core) and S2F-ratio as well as heteroskedasticity in the models residuals. 1.7988 in the Durbin-Watson statistic indicates that there are no significant autocorrelation in the residuals of the model. Since the initial model faced several difficulties, a new model was created, where the bitcoin returns are log transformed, and CPI (Core), T10YIEM and VIX are transformed into monthly changes (see table 1 for calculations) and the S2F ratio is log transformed. The new regression models are visible in table 4. 26 Variable Coefficient Std.Error T-stat Significance Intercept -0.0095 0.017 -0.545 CPI (Core) Change -2.1675 1.013 -2.139 ** T10YIEM Change 0.0783 0.035 2.268 ** VIX Change -0.0055 0.007 -0.758 LN S2F 0.0032 0.003 1.057 𝑅2 0.083 Observations 133 Table 4. All values are rounded to 4 decimal points.***,**.* indicate significance at 1%, 5% and 10% levels. As seen in table 4, using transformed values of all independent variables results in a 𝑅2 value of 0.083. Both monthly changes in core inflation and monthly changes in the 10-year breakeven inflation rate are statistically significant at both a 10% and a 5% significance level. Monthly changes in core inflation seems to have a negative effect on bitcoin returns and the monthly changes in the 10-year breakeven inflation rate seems to have a small positive effect. Both the monthly changes in volatility index and the log transformed stock-to-flow ratio are statistically insignificant at all conventional levels. Once again, testing for stationarity, multicollinearity, autocorrelation and heteroskedasticity is vital. The results are visible in table 5. Variable/element Stationarity Multicollinearity Autocorrelation Heteroskedasticity (ADF) (VIF) (DW) (BP) LN Returns Stationary - - - CPI (Core) Change Non-Stationary 1.1134 - - T10YIEM Change Stationary 1.0889 - - VIX Change Stationary 1.0646 - - LN S2F Non-Stationary 1.1229 - - Model Residuals - - 1.8617 Present Table 5. Results from tests regarding stationarity, multicollinearity, autocorrelation and heteroskedasticity in both the variables and the model's residuals. All values are rounded to 4 decimal points. 27 From table 5, it is clear that the second model faces complications in the form of non-stationarity for the variables CPI (Core) Change and LN S2F. However, multicollinearity does not seem to disturb the model anymore. A Durbin-Watson statistic of 1.8617 indicates that there is no significant autocorrelation in the model's residuals. However, heteroskedasticity is still present in the model's residuals which is not ideal for an OLS estimation regression. To solve the problem of non-stationary variables and heteroskedasticity, transformation to the LN S2F variable is made, where monthly changes of the log transformed stock-to-flow ratio is now considered. Furthermore, the monthly changes of the VIX index will be squared, considering it showcases statistical insignificance in both previous models. (for calculations see table 1). The LN S2F and VIX Change variables are therefore excluded from the next regression model, which is also the final one that covers the period 2014-2024. The last regression can be seen in table 6. Variable Coefficient Std.Error T-stat Significance Intercept 0.01 0.003 3.056 *** CPI (Core) Change -1.856 0.962 -1.928 * T10YIEM Change 0.0807 0.034 2.365 ** VIX Change^2 -0.0133 0.009 -1.401 LN S2F Change -0.1053 0.116 -0.910 𝑅2 0.09 Observations 133 Table 6. All values are rounded to 4 decimal points. ***,**.* indicate significance at 1%, 5% and 10% levels. As seen in table 6, transforming both the LN S2F and VIX Change variables, while keeping the rest of the variables from table 4, results in a 𝑅2 value of 0.09, an increase of 0.007 compared to the regression model in table 4. The monthly changes in 10-year breakeven inflation rate is statistically significant at both 10% and 5% significance level. The monthly changes in core inflation is also statistically significant at a 10% significance level, however the rest of the independent variables are not significant at any conventional level. Once again, 28 the regression result indicates that the monthly core inflation has a negative effect on bitcoin returns, with a coefficient of -1.856. Since bitcoin returns are log transformed and the core inflation is stated in monthly percentage change in decimal form, the interpretation of the coefficient should regard these transformations. The marginal effect can therefore be calculated by the formula (𝑒−0.01856 − 1) * 100, which is equal to -1.8389. This indicates that a percentage point increase in monthly core inflation will decrease bitcoin returns by roughly 1.8389 percent. Meanwhile the monthly 10-year breakeven inflation rate has a small positive effect on bitcoin return with a coefficient of 0.0807. The marginal effect of the monthly 10-year breakeven inflation rate on bitcoin returns is calculated with the formula (𝑒0.000807 − 1) * 100, which is equal to 0.0807. It indicates that a percentage point increase in the monthly 10-year breakeven inflation rate will increase bitcoin returns with 0.0807 percent. Considering the marginal effects of both the monthly changes in core inflation and the monthly changes in the 10-year breakeven inflation rate on bitcoin returns, monthly core inflation appears to have a greater magnitude of effect on bitcoin returns. The monthly core inflation showcases a weaker statistical significance compared to the monthly 10-year breakeven inflation rate, since it is only statistically significant at a 10% level. The monthly 10-year breakeven inflation rate however, is statistically significant at both a 10% and a 5% significance level. Lastly, once again testing for stationarity, multicollinearity, autocorrelation and heteroskedasticity is vital, where the results are visible in table 7. Variable/element Stationarity Multicollinearity Autocorrelation Heteroskedasticity (ADF) (VIF) (DW) (BP) LN Returns Stationary - - - CPI (Core) Change Non-Stationary 1.0132 - - T10YIEM Change Stationary 1.0707 - - VIX Change^2 Stationary 1.0528 - - LN S2F Change Stationary 1.009 - - Model Residuals Stationary - 1.8387 Not present Table 7. Results from tests regarding stationarity, multicollinearity, autocorrelation and heteroskedasticity in both the variables and the model's residuals. All values are rounded to 4 decimal points. 29 From table 7, it is evident that this model only exhibits non-stationarity in the CPI (Core) Change variable. However, the residuals of the model are stationary and therefore it is justified to use an OLS regression considering the residuals indicating a cointegrating relationship. (Engle & Granger, 1987). Moreover, no variable experiences multicollinearity and no heteroskedasticity is found in the model's residuals. Furthermore, 1.8387 in the Durbin-Watson statistic indicates that there is no significant autocorrelation in the residuals of the model. 6.2 Bitcoin as an Inflation Hedge During the COVID-19 Pandemic In table 8 the regression result for the first of January 2019 to the first of January 2023 is presented. The dependent variable is the log transformed bitcoin returns. Instead of the monthly change in the 10-year breakeven inflation rate, the monthly change in the 5-year breakeven inflation rate is utilized. The remaining independent variables are the same as in the final regression of subsection 6.1. Variable Coefficient Std.Error T-stat Significance Intercept 0.0149 0.005 2.846 *** CPI (Core) Change -2.729 1.093 -2.497 ** T5YIEM Change 0.0483 0.026 1.828 * VIX Change^2 -0.0223 0.019 -1.201 LN S2F change -0.2283 0.181 -1.263 𝑅2 0.229 Observations 49 Table 8. All values are rounded to 4 decimal points. ***,**.* indicate significance at 1%, 5% and 10% levels. As seen in table 8, the regression model for the time frame 2019-2022 has a 𝑅2 value of 0.229, which is greater than the final model presented in subsection 6.1. Monthly changes in core inflation is statistically significant at both a 10 % and a 5 % significance level and has a 30 negative coefficient of -2.729. Considering that bitcoin returns are log transformed and core inflation is stated in monthly percentage changes in decimal form, it is essential to interpret the coefficient regarding these transformations. The marginal effect is calculated by the formula (𝑒−0.02729 − 1) * 100, which is equal to -2.6921. This indicates that a percentage point increase in monthly core inflation will decrease bitcoin returns with 2.6921 percent. Furthermore, the monthly changes in 5-year breakeven inflation rate is statistically significant at a 10% significance level with a positive coefficient of 0.0483. Once again, the bitcoin returns are log transformed and the 5-year breakeven inflation rate is stated in monthly percentage change in decimal form. Therefore, the transformations should be accounted for when interpreting the coefficient. The marginal effect is calculated by the formula (𝑒0.000483 − 1) * 100, which is equal to 0.0483. This indicates that a percentage point increase in the monthly 5-year breakeven inflation rate will increase bitcoin returns with 0.0483 percent. No other independent variable in the regression is statistically significant at any conventional level. Thus, the marginal effect for these variables will not be presented. Considering the marginal effects of both monthly changes in core inflation and the 5-year breakeven inflation rate on bitcoin returns, the monthly core inflation appears to have a greater magnitude of effect on bitcoin returns compared to the monthly 5-year breakeven inflation rate. Furthermore, the monthly core inflation showcases a greater statistical significance compared to the monthly 5-year breakeven inflation rate, since it is statistically significant at both 10% and a 5% significance level. The monthly 5-year breakeven inflation is statistically significant at only a 10% significance level. After the regression, controlling for problems such as stationarity, multicollinearity, autocorrelation and heteroskedasticity is vital, since they can cause inflated 𝑅2 and coefficient values etc. In table 9, a summary of the vital tests is visualized. 31 Variable/element Stationarity Multicollinearity Autocorrelation Heteroskedasticity (ADF) (VIF) (DW) (BP) LN Returns Stationary - - - CPI (Core) Change Non-Stationary 1.0345 - - T5YIEM Change Stationary 1.3475 - - VIX Change^2 Stationary 1.227 - - LN S2F Change Stationary 1.146 - - Model Residuals Stationary - 1.847 Not present Table 9. Results from tests regarding stationarity, multicollinearity, autocorrelation and heteroskedasticity in both the variables and the model's residuals. All values are rounded to 4 decimal points. In table 9, it is evident that CPI (Core) Change is the sole variable that exhibits non-stationarity. However, the residuals of the model display a stationary relationship. Therefore, by adapting the framework constructed by Engel & Granger (1987), the OLS model is appropriate to use since the residuals indicate a cointegrating relationship. Furthermore, there is no evidence that multicollinearity and heteroskedasticity are present in the regression model. Lastly, the Durbin-Watson statistic of 1,847 indicates that there is no significant evidence of autocorrelation in the regression model. 6.3 Gold and Oil as a Hedge Against Inflation In the following subsections previous empirical research regarding gold and oil as inflation hedges are utilized to determine their hedging capabilities during both 2014-2024 and the COVID-19 pandemic. 6.3.1 Gold as an Inflation Hedge The subject of gold acting as an inflation hedge has been studied for a long time. According to a study by Valadkhani et al. (2022), gold is highly sensitive to inflation occurring in monthly U.S. inflation rates in levels above 0.55%, therefore serving as a rather effective 32 hedge during times of high inflation. Ergül and Karakaş (2024) also backed that gold could be considered as a partial hedge against inflation, a stronger safe haven asset in times of economic uncertainty. Nevertheless, the relationship between gold and inflation is not always that predictable. According to the CFA Institute (2024), the correlation between monthly changes in gold prices with the U.S. inflation rates ranged between -0.004 and 0.162, indicating a weak and unstable relation. Hence, this leads to the possibility of gold acting as a hedge in certain conditions and not always providing protection against inflation in others. 6.3.2 Oil as an Inflation Hedge Fuel prices have a cause-effect relationship with inflation. Oil prices and stock returns have found statistical significance within OECD countries-biased data, thus meaning that oil prices do influence financial markets and inflation. Then again, Breman and Storm (2023) suggested that excessive speculation in the crude oil market accounted for 24%-48% of the increase in WTI crude oil price from October 2020 to June 2022. This then led to an increase in the inflation rate of U.S. Personal Consumption Expenditures (PCE) by 0.75 to 1.5 percentage points during this time frame. Thus, the findings probe the considerable say oil has over inflation, albeit leading to its volatility in being a consistent hedge with forgivable external factors. 6.3.3 Conclusion According to empirical data & literature from 2014 to 2024, under certain circumstances, both gold and oil can be utilized as inflation hedges. Oil has a more direct effect on inflation but is also considered more volatile due to market dynamics. Gold on the other hand typically works better during times of high inflation. To maximize their investment strategies, investors looking to seek protection from inflation should take into account the features of these commodities as well as the state of the economy. 33 7. Discussion In this chapter the reader is presented with a discussion regarding the results presented in the previous chapter. Furthermore, the results are then interpreted from an economic viewpoint in order to explain the results that were found. Moreover, limitations of this thesis are presented to visualize shortcomings in both the model(s) and the data retrieved. Lastly, from the shortcomings of this thesis, suggestions for potential future research are discussed. 7.1 Results The results from our regression model during 2014-2024 indicate that bitcoin does not exhibit any inflation hedging characteristics. As seen in table 6, where the results for the time frame 2014-2024 is presented, core inflation has a strong negative effect on bitcoin returns. Accordingly, bitcoin may behave like the broader market and decrease in value when inflation rises rather than provide protection against the rising inflation. However, since core inflation is marginally significant at a 10 % significance level, we cannot confidently reject the null hypothesis at a 5 % significance level. Therefore, evidence for a robust relationship between bitcoin and inflation remains weak. This aligns with the conclusion that Choi & Shin (2022) found in their thesis. Moreover, the control variable that measures the 10-year breakeven inflation rate showcases a noticeable statistical significance, where it is statistically significant at a 5% significance level. It also shows a slight positive effect on bitcoin returns, indicating that a rise in 10-year inflation expectations affects bitcoin slightly positively. This was not a surprise since an increase in inflation expectations could indicate that fiat currencies decreased in real term valuation due to the FED printing more money during this time. Hence, bitcoin could attract investors believing its price will remain stable and even rise due to higher demand. Both the squared volatility index and log transformed stock-to-flow ratio are statistically insignificant, indicating that evidence for a robust relationship between them and bitcoin returns are weak. This further implies that bitcoin is not a safe-haven asset and does not perform better during market turmoil, which Choi & Shin (2022) also found in their thesis. Regarding the stock-to-flow ratio, Shelton (2024) found that the stock-to-flow ratio does not explain bitcoin returns alone and had a weak explanatory power in his overall model. This aligns well with the statistically insignificant stock-to-flow ratio found in our model. 34 Furthermore, our model has a relatively low 𝑅2 value which indicates that the model has a low explanatory power of the variation in bitcoin returns. This further highlights bitcoin’s complex nature and indicates that other factors not included in this thesis may have an important role in explaining variation in bitcoin returns. Other factors that may affect bitcoin returns could be regulatory shocks, investment sentiment and changes in interest rates. These may in turn improve the 𝑅2 value of our model. Our regression model for the time frame 2019-2022, which is visible in table 8, also indicates that bitcoin is not a suitable asset to use for hedging inflation. Core inflation still has a negative effect on bitcoin returns, however it is significant at a 5 % significance level. Therefore, we can confidently reject our null hypothesis. Although, considering that a rise in inflation suggests a negative change in bitcoin returns, it indicates that bitcoin may not serve as an effective inflation hedging asset, which aligns with Choi & Shin (2022) and Smales (2024) previous research. Accordingly, bitcoin may behave like the broader market and decrease in value when inflation rises rather than provide protection against the rising inflation during 2019-2022 The 5-year breakeven inflation rate used as a control variable for inflation expectations over a 5-year period is marginally statistically significant at a 10% level. It has a slight positive effect on bitcoin returns, although we cannot with confidence conclude that there is a statistically robust relationship between bitcoin and 5-year inflation expectations during 2019-2022. Considering our model not finding a robust relationship, it may indicate that the FED raised interest rates at an aggressive rate during 2022 due to the U.S economy overheating after recovering from the COVID-19 pandemic. Before 2022 it is likely that the FED printed money in order to raise economic activity in the U.S, considering the U.S economy experiencing decreasing economic activity during the COVID-19 pandemic. This may give an insight to why the coefficient is slightly positive during 2019-2022. Moreover, this second regression model aims to find bitcoin’s behaviour during times of market turmoil during the COVID-19 pandemic. However, the squared volatility index used to control for market volatility is still statistically insignificant, indicating that bitcoin does not exhibit safe-haven asset properties during times with heightened volatility. This is a contradiction from the conclusion drawn by Choi & Shin (2022), where they found that 35 heightened volatility decreases bitcoin returns. Hence, they found a relationship with a statistically significant volatility index, although it did not satisfy any of the safe-haven characteristics. Regardless, this contradiction strengthens the conclusion that bitcoin lacks safe-haven characteristics. Furthermore, the log transformed stock-to-flow ratio variable exhibits no evidence for a robust relationship with bitcoin. The variable is statistically insignificant and does not provide a favorable justification for explaining the variation in bitcoin returns. Our second model exhibits an improved 𝑅2 value compared to our first model. Unsurprisingly, this model still has a low explanatory power of the variation in bitcoin returns. Once again, other factors outside the scope of this thesis may affect bitcoin returns, which is remarkably likely considering the complex nature of bitcoin. Other factors that may affect bitcoin returns could be changes in interest rates, regulatory shocks, and investment sentiment which may in turn improve the 𝑅2 value of our model. However, given the findings in this thesis, evidence for using bitcoin as an inflation hedge has shown both an unreliable robust relationship between 2014-2024 and a negative relationship between 2019-2022, indicating that bitcoin does not serve as a reliable inflation hedging asset. Although we can’t reject our null hypothesis at a 5% significance level for the period 2014-2024, we can reject it during the period 2019-2022, where we found evidence for a robust relationship between bitcoin returns and inflation. Lastly, bitcoin is an inferior inflation hedge compared to both gold and oil. Both Valadkhani et al (2022) and Ergül and Karakaş (2024) found that gold is a favorable hedging asset when markets face economic uncertainty. Furthermore, they state that gold is rather a safe-haven asset and therefore an asset that can be utilized as a partial inflation hedging asset. We found that bitcoin lacks inflation hedging characteristics during 2019-2022, when the COVID-19 pandemic erupted. Therefore, during 2019-2022, investors who desired to protect their wealth from heightened inflation may have improved their protection by allocating to gold rather than to bitcoin. Moreover, Breman and Storm (2023) found that oil also is an appropriate inflation hedging asset during 2014-2024. They argue that oil is a driver of inflation since it has a strong 36 influence on modern society. Hence oil may act as a more consistent hedge against inflation compared to bitcoin. However, they also make a case for volatility associated with oil, which may affect the outcome of its potential inflation hedging characteristics. Nevertheless, oil also may have provided better coverage against inflation for investors since we found that bitcoin does not hedge inflation during 2014-2024 either. 7.2 Limitations While this thesis offers valuable findings into bitcoin as a possible inflation hedge, some existing limitations should be taken into consideration. One of such limitations arises from the chosen time frame from 2014 to 2024. The time period does include several crucial economic events such as the COVID-19 pandemic and the surge in inflation but can probably be seen as a bit short for these types of studies. Analysis of traditional hedge assets, such as gold and oil, tend to be in the region of a period of two decades or more. The time horizon under consideration stands the risk of leaving out long-term dynamics and structural breaks in the relationship between bitcoin and inflation (Smales, 2024). Besides, this thesis somewhat depends on month end prices taken from CoinMarketCap, and although these numbers appear sufficient, they sometimes miss small price movements and short lived trends within the market. Also, when it comes to intraday volatility, bitcoin is considered extreme, meaning it would have revealed higher hedging properties if one had used higher-frequency data (Wissmann, 2022). Hence, the choice was made to adhere to a more stable long-term dataset usually preferred in studies of traditional assets, but this could mean that a bias was injected for stable price movements. Another possible limitation would be that the linear regression models consider the relationship between bitcoin returns and inflation as being purely linear, a major simplifying assumption. Financial markets are usually complicated, with non-linearities in patterns being the rule rather than the exception, especially during times of economic stress. (Morillon & Chacon, 2022). This simplified approach may ignore the more subtle non-linear price changes in response to inflation shocks and thereby understate bitcoin’s usefulness as a hedge. Hence, this research embraces the U.S.based economic variables consisting of the Core CPI and breakeven inflation rate, which paradoxically might limit the generalization of the 37 inference onto other zones with diverse monetary policies and inflation dynamics. Meaning, not considering probable global factors like geopolitical risks, diverse monetary regimes, among others, the research scope becomes even constricted (Rodriguez & Colombo, 2025). The study may omit certain fundamentals such as the Trimmed Mean PCE of inflation, Median CPI of inflation, or alternative measures of inflation statistics. The model was estimated with these variables in the initial analysis, however, since the presence of these variables worsened the fit of the model and was less justifiable theoretically, they were removed. Hence, these variables were omitted in the final conclusions of the thesis. This thesis does deliver important findings about bitcoin's inflation hedging function, yet additional research should examine extended time frames, and diverse inflation metrics alongside nonlinear modeling approaches to evaluate bitcoin's portfolio diversification role. 38 8. Conclusion In this final chapter the reader is presented with the conclusions reached from the research conducted in this thesis. Moreover, an invitation for future research examining areas not included or areas in need of improvement in this thesis are expressed. This thesis examines if bitcoin exhibits inflation hedging characteristics in order to evaluate the use of bitcoin as an asset investors may invest in to preserve their capital from inflation. Through examining time spans between both 2014-2024 and 2019-2022, bitcoin's inflation hedging characteristics could be investigated during both a wider time frame and during the COVID-19 pandemic, where markets experienced heightened volatility and inflation. By implementing two OLS regressions, for each time span, the statistical relationship between bitcoin and inflation could be tested. Lastly, comparing bitcoin with traditional inflation hedging assets such as gold and oil further assisted with evaluating bitcoin's performance as an inflation hedging asset. Unfortunately, the results of the OLS models indicate that bitcoin does not exhibit inflation hedging characteristics. During 2014-2024, no evidence of a robust relationship between bitcoin and inflation is found and during 2019-2022 a statistically significant negative coefficient is found. This indicates that bitcoin is an asset that does not hedge inflation nor act as a safe-haven asset. Meanwhile, other empirical studies utilized in this thesis reached the conclusion that both gold and oil exhibit at least partial inflation hedging capabilities. Gold, a traditional safe-haven asset, outperformed bitcoin during 2019-2022. Furthermore, oil’s performance during 2014-2024 exhibits a more consistent relationship with inflation compared to bitcoin despite being subject to volatility, indicating a stronger capability of hedging against inflation. Thus, this thesis provides significant information for investors. Investors concerned about inflation eroding their wealth may find gold and oil a better investment for preservation rather than bitcoin. However, the research conducted in this thesis is limited to the U.S. Therefore, further research may provide different relationships for different markets. Furthermore, a low explanatory power of bitcoin returns in the OLS models used in this thesis invite further research where different independent variables and non-linear models could provide contradicting results to this thesis. 39 9. Bibliography Baur, D.G. and Lucey, B.M. (2010). ‘Is Gold a Hedge or a Safe Haven? An Analysis of Stocks, Bonds and Gold’, The Financial Review, 45(2), pp. 217–229. https://doi.org/10.1111/j.1540-6288.2010.00244.x [Accessed 29 March 2025] Bitbo. (n.d.). ‘Bitcoin Total Circulating Supply’ [online]. https://charts.bitbo.io/circulating-supply/ [Accessed 2 April 2025] Bodie, Z. (1976). ‘Common Stocks as a Hedge Against Inflation’, The Journal of Finance, 31(2), pp. 459–470. https://doi.org/10.1111/j.1540-6261.1976.tb01899.x [Accessed 29 March 2025] Breman, C. and Storm, S. (2023). ‘Betting on black gold: Oil Price Speculation and U.S. Inflation (2020–2022)’, Institute for New Economic Thinking Working Paper Series [online]. https://www.ineteconomics.org/uploads/papers/WP_208-Storm-and-Breman-Oil-price-specul ation-and-inflation.pdf [Accessed 11 May 2025] Cboe Global Markets. (n.d.).’VIX Historical Data’ [online]. https://www.cboe.com/tradable_products/vix/vix_historical_data/ [Accessed 10 April 2025] Choi, S and Shin, J. (2022). ‘Bitcoin: An Inflation Hedge but Not a Safe Haven’. Finance Research Letters, 46(B), pp. 1–8. https://www.sciencedirect.com/science/article/pii/S1544612321003810 [Accessed 20 March 2025] CoinMarketCap. (n.d.) ‘Bitcoin (BTC) Price, Charts and Market Cap’ [online]. Bitcoin price today, BTC to USD live price, marketcap and chart | CoinMarketCap [Accessed 1 April 2025] CoinMarketCap. (n.d.) ‘How Much Bitcoin Is in Circulation’ [online]. https://coinmarketcap.com/currencies/bitcoin/ [Accessed 1 April 2025] 40 Engle, R.F and Granger, C.W.J. (1987). ‘Co-integration and Error Correction: Representation, Estimation, and Testing’, Econometrica, 55(2), pp. 251-276. Co-Integration and Error Correction: Representation, Estimation, and Testing on JSTOR [Accessed 25 April 2025] Ergül, F. and Karakaş, O. (2024). ‘The Evolving Role of Gold as an Inflation Hedge: Evidence from the United States and Europe’, Acta Montanistica Slovaca, 29(4), pp. 1032-1042. https://actamont.tuke.sk/pdf/2024/n4/20joseph.pdf [Accessed 11 May 2025] (Link no longer available). Fandetti, M. (2024). ‘Gold and Inflation: An Unstable Relationship’, CFA Institute Blog [online]. https://blogs.cfainstitute.org/investor/2024/06/05/gold-and-inflation-an-unstable-relationship/ [Accessed 11 May 2025] Fernandis, N. (2024). ‘Cryptocurrency Particularly Bitcoin as a Hedge Against Inflation: Myth or Reality?’, [online]. PES University. Cryptocurrency Particularly Bitcoin as a Hedge Against Inflation: Myth or Reality? by Noyal Fernandis :: SSRN [Accessed 15 April. 2025] Flores, A. (2024). ‘Bitcoin Inflation Hedge:Can Bitcoin Protect Against Inflation?’, Paystand Blog [online]. https://www.paystand.com/blog/bitcoin-inflation-hedge [Accessed 2 April 2025] Forbes Advisor. (2024) ‘Bitcoin Halving Event 2024: Not with A Bang, But a Whimper’, [online]. https://www.forbes.com/advisor/in/investing/cryptocurrency/bitcoin-halving-2024/ [Accessed 5 April 2025] 41 Jafar, B. (2025). ‘Why Bitcoin Struggles to Match Gold's Safe Haven Status’, Financial News London [online]. https://www.fnlondon.com/articles/why-bitcoin-struggles-to-match-golds-safe-haven-status-1 2ee4858 [Accessed 2 April 2025] Kavinoky Cook LLP. (2024). ‘Bitcoin as a Hedge Against Inflation: What It Means for Institutional Investors’, [online]. https://kavinoky.com/2024/08/bitcoin-hedge-against-inflation-institutional-investors/ [Accessed 29 March 2025]. Köse, N., Yildirim, H., Ünal, E, and Boqiang, L. (2024). ‘The Bitcoin price and Bitcoin price uncertainty: Evidence of Bitcoin price volatility’, The Journal of Futures Markets, 44(4), pp. 673-695. https://onlinelibrary.wiley.com/doi/abs/10.1002/fut.22487 [Accessed 16 April 2025] Lyócsa, Š., Molnár, P., Plíhal, T. and Širaňová, M. (2020). ‘Impact of macroeconomic news, regulation and hacking exchange markets on the volatility of Bitcoin’, Journal of Economic Dynamics & Control, 119, pp. 103980. Impact of macroeconomic news, regulation and hacking exchange markets on the volatility of bitcoin - ScienceDirect [Accessed 15 April 2025] Morillon, G.T. and Chacon, G. R. (2022). ‘Dissecting the Stock-to-Flow Model for Bitcoin’, Studies in Economics and Finance, 39(3), pp. 506-523 https://ideas.repec.org/a/eme/sefpps/sef-10-2021-0409.html [Accessed 8 April] Pinchuk, M. (2023). Bitcoin Does Not Hedge Inflation. arXiv preprint. https://arxiv.org/abs/2301.10117 [Accessed 29 March 2025] Rodriguez, M. and Colombo, V. (2025). ‘Is bitcoin an inflation hedge’, Journal of Economics and Business, 133, Article 106218. https://www.sciencedirect.com/science/article/abs/pii/S0148619524000602 [Accessed 16 April 2025] 42 Royal, J. (2024). ‘Bitcoin Vs. Gold: Which Is The Better Inflation Hedge?’, Bankrate [online]. https://www.bankrate.com/investing/bitcoin-or-gold-better-inflation-hedge/ [Accessed 29 March 2025] Shah, S. (2025). ‘Crypto Can't Outshine Gold as a Vital Store of Wealth’, The Times [online]. https://www.thetimes.co.uk/article/crypto-cant-outshine-gold-as-a-vital-store-of-wealth-k22z qv0pn [Accessed 5 April 2025] Shelton, A. (2024). ‘Bitcoin Return Prediction: Is It Possible via Stock-to-Flow, Metcalfe’s Law, Technical Analysis, or Market Sentiment?’, Journal of Risk and Financial Management, 17(10), 443–466. https://www.mdpi.com/1911-8074/17/10/443 [Accessed 23 March 2025] Smales, L. A. (2024). ‘Cryptocurrency as an Alternative Inflation Hedge?’, Accounting and Finance, 64(2), pp. 1589–1611. https://onlinelibrary.wiley.com/doi/abs/10.1111/acfi.13193 [Accessed 23 March 2025] Taleb, N. N. (2021). Bitcoin, Currencies, and Fragility. arXiv preprint. https://arxiv.org/abs/2106.14204 [Accessed 29 March 2025] Tembelo, H. and Ozyesil, M. (2024). ‘Examining the Relationship between Oil Prices and Stock Returns: Evidence from OECD Countries’, International Journal of Energy Economics and Policy, 14(3), pp. 307-315. https://www.econjournals.com/index.php/ijeep/article/download/15913/7874/37417 [Accessed 11 May 2025] The Federal Reserve Bank of St.Louis. (n.d.) ‘5-Year Treasury Inflation-Indexed Security: Yield (T5YIEM)’ [online]. https://fred.stlouisfed.org/series/T5YIEM [Accessed 12 April 2025] 43 The Federal Reserve Bank of St.Louis. (n.d.) ‘10-Year Treasury Inflation-Indexed Security: Yield (T10YIEM)’ [online]. https://fred.stlouisfed.org/series/T10YIEM [Accessed 12 April 2025] U.S Bureau of Labor Statistics. (n.d.) ‘Consumer Price Index for All Urban Consumers: All Items in U.S. City Average, All Urban Consumers, Not Seasonally Adjusted’ [online]. https://data.bls.gov/timeseries/CUUR0000SA0L1E [Accessed 8 April 2025] Valadkhani, A., Nguyen, J. and Chia, M. (2022). ‘When is gold an effective hedge against inflation?’, Resources Policy, 79, Article 103009 https://www.sciencedirect.com/science/article/abs/pii/S0301420722004524 [Accessed 11 May 2025] Weil, D. (2023). ‘Is Bitcoin an Inflation Hedge? Here Are the Arguments on Both Sides’, Wall Street Journal [online]. https://prasad.dyson.cornell.edu/doc/WSJ.08Oct23.pdf (Unofficial reprint), [Accessed 28 March. 2025] Wissmann, P. (2022). Is Bitcoin a good asset for inflation hedging? A comparison between Bitcoin returns and the inflation rates of the USA, the Euro-Zone, India, Kenya, and Venezuela. Bachelor’s Thesis. University of Twente. https://essay.utwente.nl/92741/1/Wissmann_BA_BMS.pdf [Accessed 5 April 2025] 44