This thesis examines the behaviour of gold, Bitcoin, and a set of macroeconomic and financial variables between January 2008 and August 2025, a period that spans the Global Financial Crisis, the COVID-19 shock, the post-2022 inflation, and the emergence of digital assets. It asks whether gold prices can be predicted from their own history, which macroeconomic factors explain gold’s movements, and whether Bitcoin behaves as a hedge, a safe haven, or a speculative asset relative to gold and the S&P 500. These questions are addressed with three methods. An ARIMA model tests whether past gold prices forecast future ones. A Vector Error Correction Model, esti- mated at monthly frequency, relates gold to CPI, unemployment, the 10-year Treasury yield, the U.S. Dollar Index, and the S&P 500; the monthly frequency matches the release schedule of the macroeconomic series and avoids the distortions that forward- filling introduces into daily data. Finally, rolling measures, a GARCH(1,1), and a DCC-GARCH, combined with regime splits based on the official CBOE VIX, com- pare the volatility and equity correlation of Bitcoin, gold, and the S&P 500 across calm and high-stress periods. Gold’s short-run price movements are close to a random walk: one-day-ahead forecasts are accurate because tomorrow’s price is usually near today’s, but the model has little to say at longer horizons. In the monthly VECM, unemployment shocks have the largest contribution among the macro variables in the selected specification, accounting for roughly 34 percent of gold’s forecast error variance at twelve months and being associated with a positive response. CPI shocks explain about 7 percent and carry a negative average sign, contrary to the view that gold reliably hedges inflation in the short run. The dollar and yield channels are weaker still. Gold is therefore more consistent with a conditional defensive asset against broad economic deterioration than with a simple inflation hedge. Bitcoin, by contrast, is far more volatile than gold and has grown more closely tied to the S&P 500 since 2020. The evidence supports treating gold as a conditional hedge and defensive asset — one that does not satisfy the strictest safe-haven definition on the most stressed days but remains far steadier than Bitcoin — and Bitcoin as a speculative one, a distinction that has sharpened in the post-2020 period.

Gold, Bitcoin, and Macroeconomic Dynamics: An Empirical Analysis of Safe Haven, Hedging, and Speculative Asset Behavior in Financial Markets

ROY, DIPTA
2025/2026

Abstract

This thesis examines the behaviour of gold, Bitcoin, and a set of macroeconomic and financial variables between January 2008 and August 2025, a period that spans the Global Financial Crisis, the COVID-19 shock, the post-2022 inflation, and the emergence of digital assets. It asks whether gold prices can be predicted from their own history, which macroeconomic factors explain gold’s movements, and whether Bitcoin behaves as a hedge, a safe haven, or a speculative asset relative to gold and the S&P 500. These questions are addressed with three methods. An ARIMA model tests whether past gold prices forecast future ones. A Vector Error Correction Model, esti- mated at monthly frequency, relates gold to CPI, unemployment, the 10-year Treasury yield, the U.S. Dollar Index, and the S&P 500; the monthly frequency matches the release schedule of the macroeconomic series and avoids the distortions that forward- filling introduces into daily data. Finally, rolling measures, a GARCH(1,1), and a DCC-GARCH, combined with regime splits based on the official CBOE VIX, com- pare the volatility and equity correlation of Bitcoin, gold, and the S&P 500 across calm and high-stress periods. Gold’s short-run price movements are close to a random walk: one-day-ahead forecasts are accurate because tomorrow’s price is usually near today’s, but the model has little to say at longer horizons. In the monthly VECM, unemployment shocks have the largest contribution among the macro variables in the selected specification, accounting for roughly 34 percent of gold’s forecast error variance at twelve months and being associated with a positive response. CPI shocks explain about 7 percent and carry a negative average sign, contrary to the view that gold reliably hedges inflation in the short run. The dollar and yield channels are weaker still. Gold is therefore more consistent with a conditional defensive asset against broad economic deterioration than with a simple inflation hedge. Bitcoin, by contrast, is far more volatile than gold and has grown more closely tied to the S&P 500 since 2020. The evidence supports treating gold as a conditional hedge and defensive asset — one that does not satisfy the strictest safe-haven definition on the most stressed days but remains far steadier than Bitcoin — and Bitcoin as a speculative one, a distinction that has sharpened in the post-2020 period.
2025
File in questo prodotto:
File Dimensione Formato  
ROY_DIPTA_898395.pdf

accesso aperto

Dimensione 1.49 MB
Formato Adobe PDF
1.49 MB Adobe PDF Visualizza/Apri

I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14247/29007