Hedging and diversification across commodity assets

ABSTRACT

We investigate the conditional cross effects and volatility spillover between equity markets and commodity markets (oil and gold), Fama and French HML and SMB factors, volatility index (VIX) and bonds using different multivariate GARCH specifications considering the potential asymmetry and persistence behaviours. We analyse the dynamic conditional correlation between the US equity market and a set of commodity prices and risk factors to forecast the transmission of shock to the equity market firstly, and to determine and compare the optimal hedge ratios from the different models based on the hedging effectiveness of each model. Our findings suggest that all models confirm the significant returns and volatility spillovers. More importantly, we find that GO-GARCH is the best-fit model for modelling the joint dynamics of different financial variables. The results of the current study have implications for investors: (i) the equity market displays inverted dynamics with the volatility index suggesting strong evidence of diversification benefit; (ii) of the hedging assets gold appears the best hedge for the US equity market as it has a higher hedge effectiveness than oil and bonds over time; and (iii) despite these important results, a better hedge may be obtained by using well-selected firm sized and profitability-based portfolios.     

Brand and consumer engagement behaviors on Facebook brand pages: Let's have a (positive) conversation

This paper examines consumer-brand engagement on social media, focusing on the dynamic interplay between brands and consumers as well as among consumers themselves. In addition to liking and sharing behaviors, we look at what influences consumers to provide feedback to brands by commenting on brand posts, connect with other consumers by replying to their comments and agree with other consumers by liking their comments and replies. Drawing on a large longitudinal dataset from 2740 Facebook brand pages across 25 industries, a multilevel multivariate autoregressive zero-inflated negative binomial model is used to examine how brand post and response behaviors are related to specific consumer engagement behaviors. Results show that the amount of brand posts indirectly affects consumer comments through increased liking and sharing, but with diminishing returns. Liking is key as it increases positive and attenuates negative feedback, whereas sharing is associated with both positive and negative consumer-brand comments. Posting photos and videos encourages liking, with photos having increasing returns. Videos are the only format with a direct positive influence on consumer-brand comments, though with decreasing effect. Brand replies to consumer comments, even delayed ones, have little effect on negative conversations but prompt replies would be more beneficial as they bolster positive consumer comments. Moreover, our findings provide evidence for both a “positivity spiral” and “negativity spiral”, with the negativity spiral stronger than the positivity one. Taken together, these findings shed light on how marketers can stimulate positive conversations between brands and consumers on social media.     

Bootstrap rolling-window Granger causality dynamics between momentum and sentiment: implications for investors

Annals of Finance - This paper seeks to examine the unidirectional versus bidirectional Granger causality between investors’ sentiment and momentum strategies. It is based on the full sample...