Notes: The proximity between nodes and their size depends on the number of bilateral links. The links shown are significant at the 0.1% threshold. Period: 1 January 2021 - 23 February 2022.
Interconnections and contagion effects
Interconnections refer to the economic and financial links between economic players and to their exposure to common risk factors (economic, health, climate, geopolitical, etc.). These interconnections can be measured in terms of co-movements between the returns of different financial assets. Contagion effects are reflected in the increase in these market interconnections in response to extreme negative events (Forbes, 2012).
Contagion phenomena may arise from changes in investor preferences and behavioural phenomena, such as increased risk aversion, financial panics and herd behaviour, sometimes triggered by "wake-up calls" (Goldstein, 1998). This behaviour may be individually rational, when associated, among other things, with risk and liquidity constraints that lead certain investors to sell healthy assets to offset their losses in times of crisis (Dornbusch et al., 2000). This channel may be exacerbated by excessive risk-taking by some financial participants.
The study of market interconnections and contagion effects is of great importance for financial stability. The more interconnected a market is, the more exposed it is to external shocks passed on by other markets. The sudden nature of contagion is an additional risk factor, as investors may be forced to abruptly adjust their portfolios to new market conditions. The rise in interconnections therefore tends to signal an increased risk of financial market correction (Berger and Pukthuanthong, 2012, 2015).
The impact of the war in Ukraine on market interconnections
Has the Russian-Ukrainian war led to contagion effects between financial markets? To answer this question, this post compares the interconnections between financial markets (i) before the invasion of Ukraine (1 January 2021 - 23 February 2022) and (ii) during the war (24 February - 15 April 2022). The measure of interconnectedness used is based on the adjusted correlation coefficient of Forbes and Rigobon (2002), calculated following the approach put forward by Dungey et al. (2005). Unlike the standard correlation coefficient, this measure is robust to changes in volatility regimes in financial markets. For each asset class, the study covers a sample of nine advanced markets. As regards commodities, the selection aims to represent the diversity of this asset class (foodstuffs, precious metals, industrial metals, energy, textile fibres).
Before the start of the war, asset classes were relatively uncorrelated, which facilitated diversification. This observation highlights the existence of risk factors specific to each asset class evidenced by clusters (with the exception of commodities: see Chart 1 in green) and the relative absence of risk factors common to all asset classes. Two groups stand out: bonds (sovereign and corporate, in yellow and blue) and equities (in red).