Tail connectedness between artificial intelligence tokens, artificial intelligence ETFs, and traditional asset classes

Abstract

We examine extreme connectedness between the artificial intelligence (AI) tokens, artificial intelligence ETFs, and other asset classes, using the quantile VAR approach of Ando et al., (2022). We find moderate connectedness levels at mean and median quantiles, with AI ETFs (AI tokens) acting as strong (weak) net emitters (receivers) of return spillovers. Findings, confirmed by alternative testing, suggest that, during normal market conditions, AI tokens may offer utility as diversifiers for portfolios of traditional assets. However, at both lower and upper quantiles, connectedness levels increase, consistent with AI tokens and ETFs being sensitive to extreme shocks. Results suggest that AI tokens and ETFs do not diversify the risk of other assets during extreme market conditions. Finally, AI tokens, especially, may offer effective hedging at low cost for traditional assets (gold, equity, real estate, bonds, and currency), except for the oil and cryptocurrency market. Investors including AI assets in portfolios need to diligently monitor for changing market conditions.