Multivariate Systems
5 algorithms in Statistical Time Series.
In this family#
-
VAR contract
Vector autoregression: every series regressed on lags of all of them. The natural model when variables move together, and the base every impulse-response and variance-decomposition result is computed from.
fitVAR(values, lags, includeIntercept) -
Structural VAR contract
Identifies structural shocks by Cholesky decomposition of the residual covariance. The identification is recursive, which means **variable ordering is an economic assumption**: the first variable is assumed unaffected contemporaneously by the others, and reordering changes the results.
fitRecursiveSVAR(values, lags) -
VECM contract
Vector error correction with a fixed cointegrating vector. For series that wander individually but not apart: differencing them separately would throw away the long-run relationship, which is usually the thing of interest.
fitVECMFixedBeta(values, beta, differenceLags, includeIntercept) -
Impulse-Response Analysis contract
Traces how a one-off shock to one variable propagates through the system over time. The headline output of any VAR — and only interpretable given the identifying assumption that produced the impact matrix.
impulseResponses(coefficients, horizon) -
Forecast-Error Variance Decomposition contract
Attributes each variable's forecast error variance to the structural shocks, by horizon. Answers 'how much of the movement in this variable is explained by that one' — subject, again, to the identification.
forecastErrorVarianceDecomposition(coefficients, sigmaU, horizon)
What they share#
Every topic here is a record-transform, so once you have
called one the rest follow the same shape. Import paths differ only in the final segment:
import { fitVAR } from "fintech-algorithms/statistical-time-series/multivariate-systems/var";
import { fitRecursiveSVAR } from "fintech-algorithms/statistical-time-series/multivariate-systems/structural-var";Read them in the order above — the sequence is pedagogical, not alphabetical.
Where this sits#
Statistical Time Series collects 29 algorithms across 5 families. For the concept behind this family rather than the call signatures, see the concept guides.