On the co-movement of crude, gold prices and stock index in the Indian market

Abstract

The nonlinear relationship in the joint time-frequency domain has been studied for the Indian National Stock Exchange (NSE) with the international Gold price and WTI Crude Price being converted from Dollar to Indian National Rupee based on that week’s closing exchange rate. Though a good correlation was obtained during some period, but as a whole no such cointegration relation can be found out. Using the Discrete Wavelet Analysis, the data was decomposed and the presence of Granger Causal relations was tested. Unfortunately, no significant relationships are being found. We then studied the Wavelet Coherence of the two pairs, namely NSE-Nifty & Gold and NSE-Nifty & Crude. For different frequencies, the coherence between the pairs have been studied. At lower frequencies, some relatively good coherence have been found. In this paper, we report for the first time the co-movements between Crude Oil, Gold and Indian Stock Market Index using Wavelet Analysis (both Discrete and Continuous), a technique which is most sophisticated and recent in market analysis. Thus, for long-term traders they can include gold and/or crude in their portfolio along with NSE-Nifty index in order to decrease the risk (volatility) of the portfolio for the Indian Market. But for short-term traders, it will not be effective, not to include all the three in their portfolio.

An analysis of crude oil prices in the last decade (2011-2020): With deep learning approach

“…Kumar [38] found non linear relationships between Crude Oil and Gold prices from an Indian perspective. We also studied in past the co-movement of Crude Oil and Gold prices with Indian stock index [39]. Wang and Chueh [40] found out that in the short term, Gold prices and Crude oil prices have a positive influence on each other; while in long term the interest rates have a negative influence on future Gold prices and a positive influence on future Crude Oil prices.…”

Technical Note: Improved partial wavelet coherency for understanding scale-specific and localized bivariate relationships in geosciences

“…The new method may benefit from a better statistical significance testing method. Options for multiple testing can be the Bonferroni adjusted p test (Westfall and Young, 1993) or false discovery rate (Abramovich and Benjamini, 1996;Shen et al, 2002), which is less stringent than the former. The AR(1) model was used to generate noise series for testing the confidence level of PWC.…”