R: Diebold–Yılmaz connectedness (using the 'connectedness' package)

title: Connectedness

slug: connectedness

Definition

A framework, pioneered by Diebold and Yılmaz (2012, 2014), for measuring how shocks transmit across variables in a system using forecast-error variance decompositions from a vector autoregression. The total connectedness index captures system-wide interdependence, while directional measures show how much each variable transmits to and receives from others.

Why It Matters

In an interconnected economy or financial system, understanding the direction and magnitude of spillovers is essential for risk management and policy. The connectedness approach quantifies systemic risk, reveals which variables are dominant transmitters or vulnerable receivers of shocks, and tracks how these roles evolve over time. This makes it a key tool for monitoring contagion, assessing financial stability, and designing macroprudential policies.

Example

A central bank studies spillovers among equity markets in the US, Eurozone, and Asia. A variance decomposition from a VAR shows that US equity shocks explain 35% of the forecast-error variance of Asian markets at the 10-day horizon, while Asian shocks explain only 8% of US variance. The total connectedness index summarizes this network into a single number that can be tracked daily using a rolling-window estimation.

```r

library(connectedness)

fit <- VAR(data, p = 4)

spillover(fit, n.ahead = 10, no.corr = FALSE)

```

Related Terms

• FEVD
• Cointegration
• Cross-Quantilogram

Software Notes

R: Use the connectedness package for full Diebold–Yılmaz spillover tables and rolling-window analysis. Python: The connectedness library on PyPI offers similar functionality. MATLAB: Diebold and Yılmaz provide replication code on their websites.

Contact Us for Support → /contact/