The Influence of Gold Market on Bitcoin Prices : Is there a correlation?

Sammanfattning: Background: This paper analyses the influence of fluctuation in gold market on bitcoin prices. Based on previous studies, in present market conditions, volatility in gold prices have caused price changes in several other major assets in the market, such as crude oil. Gold fluctuations are likely to stimulate uncertainty in some other major assets. As bitcoin is becoming an alternative tool to hedge against inflation likewise to gold, the degree of uncertainty in bitcoin market is relatively high. Therefore, the study of causal relationship between gold and bitcoin markets has become appropriate since bitcoin has tremendous growth in its returns and shares many similarities with gold. Thereupon, this study reveals the evidence of Granger causality regression in different time spans to understand the relationship between gold and bitcoin. This relationship is beneficial to study since Granger causality hypothesis acknowledges whether gold’s historical prices are useful for forecasting the bitcoin market. Purpose: This study aims to analyze the relationship between gold and bitcoin market during an 8-year period from 2014 and 2022. Throughout this period, time spans which involves financial crises have been separated from the data set and tested separately to determine if there is a constant relationship between the variables. Through this, it has been intended to find the Granger causality link between gold and bitcoin market to see whether one is leading another one. Identifying the Granger causality correlation helps analyzing the patterns of correlation by using the empirical datasets, and to determine the strength of the Granger causal relationship’s nature between gold and bitcoin. Since the correlation itself does not explain why or how, but only if both markets move together, the Granger causality correlation between gold and bitcoin is the quantification of the impact that gold market performance has on bitcoin’s future price performance. Method: Since the collected data is time-series data, Augmented Dickey-Fuller tests have been conducted initially to the chosen tests. Following the results from ADF tests, Spearman’s Rho, iand Johansen’s Cointegration tests have been utilized to determine the long-term correlation between variables. Thereafter, Toda & Yamamoto and Dolado & Lütkepohl Granger Causality (TYDL-GC) method has been used to analyze the Granger causality link between the variables. Conclusion: The results of this study indicates that (i) no statistically significant correlation between gold and bitcoin market has been found according to the Spearman’s Rho test results, (ii) no long-term relationship has been found between gold and bitcoin according to cointegration test, (iii) gold does Granger Cause bitcoin prices. The evidence of causality link is unilateral from gold towards bitcoin market. Furthermore, it was observed that the Granger causality link weakens in short term and is not constant over time. The results fail to support the semi strong Efficient Market Hypothesis form. Thus, gold and bitcoin’s markets are efficient in the weak form but inefficient in the semi strong form. Since Granger causality has been found from gold towards bitcoin, one can construct a prediction model for bitcoin by using gold’s historical prices.